15 results about "Decomposition problem" patented technology

The invention provides a complex question and answer method and system based on adaptive task deconstruction and multi-modal evidence aggregation, belongs to the technical field of natural language processing, and designs a dynamic Few-shot prompt construction method based on dependency syntax fingerprints to ensure that a prompt template is matched with a question structure; the invention discloses a dynamic problem deconstruction method based on confidence evaluation and auto-reflection. The method comprises the following steps: recursively decomposing a problem tree by using a large language model; converting the problem tree into a standardized linear task execution sequence by a problem tree context dependence specification and task sequence generation method; obtaining a high-correlation evidence set of each task based on an evidence generation method of two-way recall and cross encoder rearrangement; and the task sequence is reasoned and dynamically optimized by a question answer extraction method based on double-strategy aggregation reasoning. According to the method, the accurate complex question and answer result can be provided on the premise of ensuring the question disassembling quality, restraining error propagation and comprehensively recalling evidences.

The invention provides an anomaly detection method and device for image data and a storage medium, and relates to the technical field of image processing, and the method comprises the steps: carrying out the feature extraction of the image data, fusing the spectrum and spatial features to obtain a joint feature matrix, and inputting an anomaly detection model; the model adopts an alternating direction multiplier algorithm to solve a low-rank sparse decomposition problem, and an objective function comprises a data fidelity item, a regularization item and a waveband weight item; the regularization item comprises a low-rank constraint and a sparse constraint, and the wave band weight item acts on the low-rank constraint in a weighting form; updating a background low-rank tensor, an abnormal sparse tensor, a Lagrange multiplier, a sparse constraint weight, a wave band weight item and penalty parameters of an algorithm by adopting an iteration mode in a solving process; and repeating iteration until a preset termination condition is reached, calculating an abnormal score graph pixel by pixel based on the abnormal sparse tensor, and comparing to determine an abnormal target. The problem that an abnormal target is difficult to accurately recognize in a complex scene can be solved, and detection precision and efficiency are improved.

Quantum optimization with Rydberg atom arrays is provided. In particular, methods are provided for solving combinatorial graph optimization problems, constraint satisfaction problems, maximum independent set problems, algebraic problems, and factoring.

The invention discloses a multi-table fine-grained retrieval method and device based on large model decomposition reasoning. The method comprises the following steps: firstly, prompting a table cell value and field description of each field of a table to a large model to obtain a field-level semantic annotation; then, relevant table fields are positioned through an analysis-mapping-filling strategy, a large model is used for understanding and decomposing problems, and key information such as entities and conditions is extracted; a field set required by the question is screened out in combination with table structure information; then, based on a field retrieval result, analyzing a value range of the field and further positioning related cells; and finally, integrating the retrieved fields and cells to generate a fine-grained target sub-table which is used for answering user questions. Compared with whole-table-level retrieval, the method only captures required information like human reading, significantly reduces redundant interference, improves precision ratio and response speed, and can be widely applied to medical treatment, finance, complex database management and other scenes with strict requirements for accurate structured data retrieval.

The embodiment of the invention provides a multiphase pipe flow multi-parameter optimization method and system based on a genetic algorithm, and belongs to the field of multiphase fluid mechanics and oil and gas pipeline engineering. The method comprises the following steps: constructing a corresponding multi-objective function model based on a decomposition problem of a target multi-phase pipe flow problem; wherein the decomposition problem comprises any one or more of a liquid holdup accuracy problem, a friction resistance coefficient accuracy problem and an on-way length accuracy problem; calling a genetic algorithm based on a multi-objective function model, and carrying out individual coding and group initialization operation; executing the genetic algorithm, and executing optimization solution on the multi-objective function model; and repeating iteration until the optimized multiphase pipe flow calculation model meets the target precision, and outputting an individual with the highest fitness as a final optimization result. According to the scheme, the problems that a traditional method cannot give consideration to multi-objective optimization and the global search capability is insufficient are solved, and reliable technical support is provided for simulation and control of oil and gas field multiphase flow.

An acceleration method leveraging distributed microservices for large-scale optimization problems. The method is applicable to a generalized Benders decomposition problem, a least-cost planning and a reliability analysis. The invention decomposes the optimization problem into a set of optimization subproblems which are solved iteratively. Each subproblem carries out the following tasks: data loading; data preprocessing; optimization problem creation; optimization problem resolution; results validation; and results storing and sending. The invention is implemented asynchronously and uses distributed incremental acceleration. Instead of creating and solving a different mathematical problem on each iteration, the same problem is solved on each iteration with minor modifications. This allows the leveraging of microservices-driven distributed computing with a sufficient quantity of processors leading to a dramatic improvement in performance and accuracy.

The present invention provides a homomorphic encryption federated learning method based on the polynomial decomposition problem, comprising: a server broadcasting global model parameters to each client; after receiving the global model parameters, the client uses local data to train the global model as its own local model and sends it to a designated trusted third-party encryption module; the third-party encryption module converts the local model parameters into a parameter matrix of the same type; initializes the element positions in the parameter matrix to polynomial fractions; and sends the polynomial fractions of all clients to the server; after receiving the polynomial fractions from the third-party encryption module, the server uses a definition method to calculate the corresponding abnormal integral and gradually restores the aggregated model. The present invention can further prevent privacy leakage and enhance the privacy security of the federated learning framework; the present invention has the advantage of being able to resist poisoning attacks on client model parameters, as well as inference and inversion attacks on the model.

The present invention discloses a method for fast matching of load characteristics based on an AdaBoost model, and relates to the field of power load monitoring. At present, mathematical optimization and pattern recognition are two major types of solution methods for load identification and decomposition problems, but the solution efficiency of mathematical optimization is low and the accuracy of load identification algorithm is low. To this end, the present invention proposes a method for fast matching of load characteristics based on an AdaBoost model, which uses the graphic features in the load V-I curve trajectory as identification samples, defines identification labels for each load, selects appropriate weak identifiers, generates enhanced strong identifiers, and finally realizes load identification. The use of the AdaBoost identifier can extract important training data features, reduce the dimension of training, speed up training, and improve the accuracy of training. The proposed method can quickly, efficiently and accurately match load characteristics, and has good economic benefits and practical value.

A reweighted step-by-step regularized bearing fault diagnosis method based on sparse decomposition relates to bearing fault diagnosis. The method includes the following steps: 1) establishing a sparse decomposition problem model: by introducing a suitable penalty function to replace the 0 norm, and by means of regularization, transforming it into an unconstrained optimization problem; 2) constructing a substitute function so that the substitute function has an optimal solution close to the objective function; 3) optimizing the substitute function, by iteratively constructing and reducing the substitute function, the time index parameter gradually converges to the position where the actual impact occurs, and the sparse coefficient gradually approaches the true value; 4) updating the regularization parameter and eliminating redundant atoms, when the convergence error is less than the set convergence threshold, ending the iteration, and completing the reweighted step-by-step regularized bearing fault diagnosis. It can have both strong sparse promotion ability and high fitting accuracy, and has higher signal reconstruction accuracy, so as to provide more accurate bearing fault diagnosis results.

The application provides a database native time series decomposition method and system, which comprises the following steps: converting a database native time series decomposition problem into a linear equation solving problem based on a preset single round period trend decomposition algorithm, and determining a parameter matrix of the linear equation; performing forward substitution calculation on time series data of a single page in the database based on a preset decomposition precalculation algorithm according to the parameter matrix, and obtaining a precalculation result; the calculation parameters of the decomposition precalculation algorithm correspond to the timestamps of the time series data of the single page; performing recalculation on the precalculation results of the first and last parts of a query range based on a preset recalculation algorithm according to the query range, and obtaining the recalculation result; and performing backward substitution calculation on the recalculation result based on a preset query decomposition algorithm, and obtaining a decomposition result of the database native time series. The application can quickly respond to the decomposition query request of a user, thereby significantly improving the decomposition result query efficiency.

An acceleration method leveraging distributed microservices for large-scale optimization problems. The method is applicable to a generalized Benders decomposition problem, a least-cost planning and a reliability analysis. The invention decomposes the optimization problem into a set of optimization subproblems which are solved iteratively. Each subproblem carries out the following tasks: data loading; data preprocessing; optimization problem creation; optimization problem resolution; results validation; and results storing and sending. The invention is implemented asynchronously and uses distributed incremental acceleration. Instead of creating and solving a different mathematical problem on each iteration, the same problem is solved on each iteration with minor modifications. This allows the leveraging of microservices-driven distributed computing with a sufficient quantity of processors leading to a dramatic improvement in performance and accuracy.

The present invention discloses an adaptive variational mode decomposition method for rolling bearing fault feature extraction, including: (1) building a bearing data acquisition platform to collect the bearing fault vibration signal x(t), and initializing the number of intrinsic mode components IMF, that is, the parameter k = 1; (2) setting k = k + 1, performing variational mode decomposition on the collected vibration signal x(t), and respectively calculating the permutation entropy PE of the k-th and the (k-1)-th IMF components, and their values are respectively denoted as PE k and PE k‑1 , if |PE k - PE k‑1 | is greater than the threshold u1, then go to (2), otherwise, set k = k - 1, and perform VMD decomposition on the signal x(t); (3) if the value of k is less than 2, go to (5); (4) calculate the Pearson correlation coefficient e between adjacent components among the k IMF components, if there is an e value greater than the threshold u2, then set k = k - 1, perform VMD decomposition on the collected vibration signal x(t), and go to (3); (5) output k discrete IMF feature components. The present invention can adaptively determine the number of decomposition modes, effectively suppress the over-decomposition problem occurring in the decomposition process, and further improve the performance of bearing fault signal mode decomposition.

The present application proposes a method, apparatus, equipment and medium for real-time small motion target detection using spatiotemporal parallel coupled robust principal component analysis, including: constructing a spatiotemporal data matrix based on the image data to be detected; constructing the coarse-grained sparse decomposition problem based on the rough low-rank matrix to be solved and the extreme sparse matrix to be solved, and constructing the fine-grained sparse decomposition problem based on the extreme low-rank matrix to be solved and the rough sparse matrix to be solved; solving the coarse-grained sparse decomposition problem and the fine-grained sparse decomposition problem respectively based on the Lagrange multiplier algorithm to obtain a rough low-rank matrix and an extreme low-rank matrix; using an edge continuity clustering detection algorithm to process the rough low-rank matrix and the extreme low-rank matrix to obtain small motion target detection results. The present invention can directly abandon the background modeling process, can directly detect targets from the sparse matrix, and improve the efficiency of target detection.

The invention discloses an index comprehensive decomposition method and system for product technical support demand conversion, and belongs to the technical field of equipment maintenance, and the method comprises the steps: obtaining technical support index decomposition parameters of a product technical support index decomposition problem; based on the technical support index decomposition parameters, establishing a comprehensive target decomposition model; and solving the established comprehensive target decomposition model by using an adaptive genetic algorithm, and outputting optimized technical support index decomposition parameters. According to the method, the comprehensive target decomposition model is constructed, and the built comprehensive target decomposition model is solved by using the adaptive genetic algorithm, so that the objectivity and the accuracy of comprehensive decomposition of indexes of technical support demand conversion are improved.

The application provides an anomaly detection method, device and storage medium for image data, and relates to the technical field of image processing. The method comprises the following steps: performing feature extraction on image data, fusing spectral and spatial features to obtain a joint feature matrix, and inputting the joint feature matrix into an anomaly detection model; the model uses an alternating direction multiplier algorithm to solve a low-rank sparse decomposition problem, and a target function comprises a data fidelity term, a regularization term and a band weight term; the regularization term comprises a low-rank constraint and a sparse constraint, and the band weight term acts on the low-rank constraint in a weighted form; in the solving process, an iteration method is used to update a background low-rank tensor, an anomaly sparse tensor and a Lagrange multiplier, as well as a sparse constraint weight, a band weight term and a penalty parameter of the algorithm; the iteration is repeated until a preset termination condition is reached, an anomaly score map is calculated based on the anomaly sparse tensor, and an anomaly target is determined by comparison. The application can solve the problem that it is difficult to accurately identify an anomaly target in a complex scene, and improve detection accuracy and efficiency.