67 results about "Data reorganization" patented technology

The invention relates to the field of semiconductors, in particular to an answer generation method and device based on a semiconductor knowledge base and a medium. The method comprises the steps of performing data association on structured knowledge according to a semiconductor domain knowledge structure to construct a class of knowledge bases; performing data recombination on the unstructured knowledge according to the layout so as to construct a second-class knowledge base; a query question is received, retrieval is executed in the first-class knowledge base and the second-class knowledge base based on a preset retrieval model, and a plurality of related knowledge fragments are obtained; determining a first fusion weight between the knowledge fragments of the same knowledge type based on the relevancy between each knowledge fragment and the query question and the update time of each knowledge fragment; determining an optimal knowledge type according to the group feature information, and determining a second fusion weight between different knowledge types according to the optimal knowledge type; and based on the first fusion weight and the second fusion weight, obtaining a reference answer according to the plurality of knowledge fragments. And the requirements of the semiconductor field on information comprehensiveness and answer accuracy are considered.

The invention discloses a method and equipment for compressing multi-dimensional data based on column storage and self-adaption, and the method comprises the steps: recombining original multi-dimensional data into a column data set structure, and organizing data block storage for the column data set structure according to a hierarchical structure; each data block stores observation grid data of a single variable of multi-dimensional data at a certain moment so as to support column type storage and self-adaptive compression; and carrying out real-time statistical analysis on characteristics or parameters of variables in the data blocks, and dynamically selecting and configuring a filter and parameters thereof to compress the multi-dimensional data. The data of the same variable are stored together through column storage, so that the data access efficiency is greatly improved, especially during variable-level analysis. Meanwhile, the self-adaptive compression technology can dynamically select a compression mode according to the distribution characteristics of the data, so that the compression rate of the data is effectively improved, and the occupied storage space is reduced. The processing speed of the meteorological grid point data can be obviously improved, and the storage cost is reduced.

The invention discloses a data processing method and device, electronic equipment and a computer storage medium, and relates to the technical field of data acquisition, the method comprises the following steps: acquiring an acquisition data packet sent by each data acquisition module, and analyzing according to the acquisition data packet to obtain successfully analyzed data, the successfully analyzed data at least comprising an acquisition node identifier; determining a memory address identifier of the successfully analyzed data according to the acquisition node identifier and a preset Hash mapping table, and after the successfully analyzed data is stored in a memory buffer area mapped by the memory address identifier, performing data quality verification according to the successfully analyzed data to obtain a quality verification result; after the data quality verification is completed, data reorganization is carried out according to the successfully analyzed data to obtain reorganized data, data restoration is carried out on the reorganized data according to the quality verification result to obtain final output data, and the method and device aim at improving the processing capacity of large-scale data collection so as to ensure the accuracy and timeliness of data collection.

The invention provides a memory access system oriented to vector and matrix calculation, the system comprises a vector memory access unit, a matrix memory access unit, a vector register group and a matrix register group, the vector memory access unit is connected with a memory interface and the vector register group, reads elements of a one-dimensional data structure or a two-dimensional data structure from a memory, and stores the elements of the one-dimensional data structure or the two-dimensional data structure; vector data are generated through data reorganization operation, a matrix access unit is connected with a memory interface and a matrix register set, matrix block data of a two-dimensional data structure are read from a memory together with the matrix access unit, matrix data are generated after data reorganization, and the matrix data are broadcasted to one or more computing units according to rows or columns. And performing calculation on the matrix data and the vector data. In the memory access system, the vector memory access unit and the matrix memory access unit can load data in parallel, and the utilization rate is improved through a plurality of computing units, so that the problems of low memory access efficiency and low memory bandwidth utilization rate are solved.

The invention discloses a transformer substation work ticket knowledge graph construction method based on a large language model, and belongs to the field of energy power and artificial intelligence. The method comprises the following steps: based on standardized electrical drawing annotation data, generating an independent knowledge graph through data recombination preprocessing, rough knowledge graph construction and attribute supplementary reasoning; all the independent knowledge maps are divided into a framework layer map and a detail layer map according to voltage grades, and full map fusion is completed by adopting a hierarchical fusion strategy; carrying out breakpoint detection by using an isolated point detection algorithm based on degree centrality, carrying out connectivity analysis through depth-first search, and re-fusing multiple non-connected sub-graphs; and screening abnormal nodes based on an electrical connection rule, inputting the abnormal nodes into a large language model to automatically supplement Agents to generate correction suggestions, and completing map optimization. Through cooperation of the large model Agent and multiple algorithms, the electrical data structuring efficiency and accuracy are improved, and knowledge network support is provided for automatic operation and fault diagnosis of the transformer substation.

The application discloses a file unshelling method and electronic equipment, and relates to the technical field of computer security, and comprises the following steps: inserting a probe at a target system call of a kernel, and injecting the probe into a target program; identifying a suspiciously shelled process through the target program; monitoring a memory permission change behavior performed by the suspiciously shelled process through the target program, and counting a permission change frequency of a memory region, an entropy value of the memory region and a call chain depth; when it is confirmed that the suspiciously shelled process is in an unshelling stage according to the permission change frequency, the entropy value and the call chain depth, writing data decrypted by the suspiciously shelled process into a target buffer through the target program; reading the data from the target buffer in a user mode, and recombining the data into a memory image according to a base address and a dirty page bitmap and writing the memory image back to a file. The method can avoid I / O bottlenecks caused by full memory dumping, reduce invalid data in the user mode, and reduce CPU occupation and memory bandwidth.

The present application provides a kind of data distribution device, data distribution system and data distribution program.Data distribution device (2) has: update data acquisition unit (9a), obtains the multiple update data corresponding to multiple write object electronic control device;Reorganization information acquisition unit (9b), obtains the reorganization information for by main device parallel execution to multiple write object electronic control device transfer update data transfer processing;Data packet generation unit (9c), multiple update data are segmented based on reorganization information to generate segmented data, and the segmented data of multiple update data generated are packed to generate multiple data packets;And distribution control unit (9e), multiple data packets are distributed to main device.

The application discloses a kind of mask protection circuit, device, chip and electronic equipment, it is related to information security technical field, in the circuit: timing control module is used to generate timing control signal based on the input system clock signal, and timing control signal is transmitted to routing encoding module;Data buffer module is used to receive the first bit width external input data, and after converting external input data into the second bit width intermediate input data, intermediate input data is transmitted to routing encoding module;Routing encoding module is used to recombine intermediate input data, generate the third bit width target input data, and according to timing control signal, target input data is distributed to compression accumulation tree module;Compression accumulation tree module includes multiple compression accumulation tree units, each compression accumulation tree unit is used to operate target input data based on double-track logic, and outputs the fourth bit width share pair operation result.This application can effectively reduce circuit area and reduce calculation delay.

The invention provides an intrusion detection device and method, electronic equipment and a storage medium, and belongs to the technical field of network security, and the method comprises the steps that a data collection module obtains a data frame of a network interface through multiple buffer areas, and sends the data frame to a protocol analysis module; the multiple buffer areas are a plurality of data buffer areas working alternately; the protocol analysis module carries out data recombination processing on the data frame, generates a to-be-detected data packet, and sends the to-be-detected data packet to the anomaly detection module; the anomaly detection module performs anomaly feature matching on the to-be-detected data packet based on the executable detection object to generate an intrusion detection result; the executable detection object is obtained by analyzing the rule description file based on the rule loading module; according to the method and the device, data acquisition is performed by adopting a plurality of alternately working data buffer areas, parallelization and buffering of data read-write operation are realized, and a data recombination and rule pre-analysis technology is combined, so that the problem of packet loss when the network flow is relatively large can be effectively solved, and the security of the system is improved.

The invention discloses a radiotherapy information management system based on multiple data sources and a construction method thereof, and relates to the technical field of medical information, and the system comprises a multi-source data obtaining module which is used for obtaining discrete data of each heterogeneous system; the multi-source data intelligent aggregation engine is used for recombining discrete data into format data and developing a dynamic form generator; the decision support rule base is internally provided with verification rules and is used for carrying out real-time verification and abnormal intervention on the format data; and the three-screen linkage workbench is used for integrating visual linkage browsing of patient information, image data and dose planning. According to the invention, full-link automatic integration and multi-system collaborative optimization of patient data are realized, so that the accuracy, efficiency and expandability of a radiotherapy process are improved.

The application discloses a substation work ticket knowledge graph construction method based on a large language model, and belongs to the field of energy power and artificial intelligence. The method comprises the following steps: based on standardized electrical drawing annotation data, independent knowledge graphs are generated through data reorganization preprocessing, rough knowledge graph construction and attribute supplement reasoning; all independent knowledge graphs are divided into framework layer graphs and detail layer graphs according to voltage levels, and hierarchical fusion strategies are adopted to complete full graph fusion; breakpoint detection is carried out by using an isolated point detection algorithm based on degree centrality, connectivity analysis is carried out through depth-first search, and non-connected subgraphs are re-fused; abnormal nodes are screened based on electrical connection rules, a large language model is input to automatically fill in Agent to generate correction suggestions, and graph optimization is completed. Through the cooperation of the large model Agent and multiple algorithms, the application improves the efficiency and accuracy of electrical data structuring, and provides knowledge network support for substation automation operation and fault diagnosis.

The invention provides a data processing method of a column storage database and a related product. The data processing method comprises the following steps: performing a hash strategy-based set operation on a column storage database to obtain an intermediate result; under the condition that the capacity of the intermediate result exceeds a preset memory threshold value, creating a temporary file based on a line storage structure in a disk; variable-length data columns are obtained from the intermediate result column by column, each variable-length data column comprises multiple rows of variable-length data, and each row of variable-length data comprises a data length value, a data numerical value and a flag bit; performing data recombination on each row of variable-length data in the obtained variable-length data column so as to reduce the data length of the variable-length data; and writing the variable-length data column after data recombination into a temporary file. According to the data processing method, the data length of the variable-length data is reduced, the space occupation of the variable-length data column of the intermediate result of the set operation based on the Hash strategy during disk falling is reduced, and the disk falling performance of the system is improved.

The application discloses a DMA accessing method and device supporting online data reorganization, and relates to the technical field of data processing, which comprises the following steps: when a DMA engine receives a DMA transmission instruction, it is judged whether there is a data reorganization instruction in the DMA transmission instruction; if yes, the data transmission direction and data reorganization parameter information are obtained; the DMA engine analyzes and splits the received DMA instruction into an access request and sends the access request to an upstream storage in the transmission direction; when the DMA engine receives access response data sent by the upstream storage, the response data is reorganized and stored in a data buffer in the DMA engine; when the data buffer collects all data in the current DMA transmission instruction, the data is forwarded to a downstream storage in a reorganized format according to the data reorganization parameters in the DMA transmission instruction. The application supports online data reorganization, makes the data reorganization process transparent to the operation core, provides convenience and flexibility for programmers, and reduces the implementation cost and power consumption of data format conversion.

The application provides a Redis high-hot data migration method, which comprises the following steps: counting original request information sent from a host to one or more slaves corresponding to the host in a Redis cluster, asynchronously analyzing the original request information and pushing the analyzed request information; recombining the pushed analyzed request information to count the access pressure of each data in the one or more slaves; and obtaining the load condition of the host, and when it is determined that the load of the host meets the predetermined standard of the host, migrating the data with the top N access pressures in all data in all slaves to a backup slave, wherein N is a positive integer. The application can realize automatic migration of high-hot data in a Redis cluster, and avoid the slave crash and cluster crash caused by all access quantities hitting the same slave in the prior art.

An iCloud data decryption and extraction system and method, relating to the field of data security technology, includes: verifying user compliance authorization and initializing the decryption environment; obtaining the user's root key based on a temporary access token and the user's iCloud account compliance authorization credentials; deriving session keys based on the user's root key and iCloud data storage structure information to obtain temporary session keys for various data types; performing parallel multi-threaded scheduling on the data to be extracted; matching decryption algorithms for various types of data downloaded to the decryption end and performing data decryption operations based on the temporary session keys; performing data download integrity verification and data decryption integrity verification; when both verifications pass, performing data reconstruction operations and reconstructed integrity verification; and outputting and securely destroying the extracted complete data types, significantly improving data extraction efficiency and data decryption compatibility and integrity.

This invention provides a training method for a distribution network voltage prediction model and a method for predicting distribution network voltage. The method includes: acquiring a training set; training a preset machine learning model based on the training set to obtain a first machine learning model; when the quality of the sample data in the training set does not meet a first preset requirement or the first prediction result of the first machine learning model does not meet a second preset requirement, performing data correction and data recombination processing on the training set based on the probability density function of the training set to obtain a recombined training set; using the recombined training set as a new training set, re-executing the steps of obtaining the first machine learning model based on the training set and subsequent steps until the quality of the sample data meets the first preset requirement and the first prediction result meets the second preset requirement, and then using the first machine learning model as the target distribution network voltage prediction model. This invention can train a high-precision distribution network voltage prediction model, thereby achieving accurate voltage prediction at each node in the distribution network.

The invention discloses an industrial data endogenous security method, system and equipment based on a physical unclonable function and a dynamic entropy pool. The method comprises the following steps: generating an unclonable device root key at an industrial device end by using a PUF technology; constructing a dynamic entropy pool fusing the environment, the data and the service entropy; deriving a one-time session key based on the device root key and the dynamic entropy value; encrypting the industrial data by adopting a post quantum cryptography and symmetric cryptography hybrid mechanism; and carrying out dynamic slicing and confusion processing on the ciphertext to generate a data packet containing a mixed slice and an encryption mapping table. And at the access end, authority verification is performed through the block chain smart contract, a temporary access token is issued, and data recombination and decryption are completed by using the token. According to the invention, internal generation of the security slave device is realized, a full-link, self-adaptive and anti-quantum computing security guarantee from slave device identity and data generation to storage access is provided, and the method is especially suitable for a high-value and high-sensitivity industrial manufacturing environment.

The invention relates to a shield load prediction method based on deep learning. The method comprises the following steps: firstly, carrying out wavelet noise reduction on shield construction historical load, geology and tunnel structure data; secondly, recombining data, inputting the data into a CNN model fused with a multi-attention mechanism, performing pooling compression in a compression layer, calculating attention weight, updating an input matrix in a reconstruction layer, extracting a spatial feature vector through convolution operation, inputting the vector into a TCN model fused with the attention mechanism, extracting and optimizing time features through causal convolution and expansion convolution, and obtaining a TCN model fused with the attention mechanism; optimized time dependence characteristics are obtained, and shield thrust and torque predicted values are calculated based on the characteristics. According to the method, the defects that a traditional method is poor in universality, low in efficiency, difficult to denoise and incapable of synchronously capturing the spatial-temporal characteristics are overcome, the load prediction precision under the complex geological condition is improved, and support is provided for construction optimization and safety.

The invention relates to the technical field of intelligent wearable equipment, and discloses an AI glasses data low-power-consumption storage management method and system, and the method comprises the steps: obtaining a multi-source signal data unit which is collected by AI glasses and is provided with a block identification code; performing transmission processing to obtain cloud data; performing data recombination according to the cloud data to obtain an intermediate data stream; performing integrity verification according to the intermediate data stream to obtain a complete environment data stream; performing multi-level distribution of target data according to the complete environment data stream to obtain a multi-level distribution state, and performing distribution verification on the multi-level distribution state to obtain initial response distribution; performing consistency verification according to the initial response distribution to obtain a consistency response set; and carrying out storage configuration optimization according to the consistency response set to obtain a final data storage scheme. According to the method, efficient collaboration among multi-layer architectures can be realized, and the reliability of data transmission and the storage management efficiency are improved.

The invention discloses a cross-domain data message exchange method and device, equipment and a medium, and relates to the technical field of electric data information processing. The method comprises the following steps: a sending terminal of a first network domain converts a to-be-transmitted data message into a binary data stream; generating a transmission start control code; the receiving terminal of the second network domain scans and analyzes the transmission start control code to generate a ready control code; the sending terminal scans and analyzes the ready control code, enters a data block transmission cycle, and codes one or more data blocks of the binary data stream into data codes in sequence; the receiving terminal scans the data code and generates a receiving success control code; the sending terminal scans and analyzes the receiving success control code, and sends the next data block or ends transmission; after transmission of all the data blocks is completed, the sending terminal generates a sending end control code; the receiving terminal scans and sends the end control code, carries out data recombination and integrity verification, and generates a transmission end control code; and the sending terminal scans the transmission ending control code to end the transmission session.

This application provides a data segment reassembly method, apparatus, storage medium, and electronic device. The data segment reassembly method includes: a splicing step: obtaining a first data segment from a data segment queue and determining the data information of the first data segment; determining, based on the data information, whether a second data segment with an adjacency relationship to the first data segment exists in the reassembly context; if a second data segment exists, splicing the first data segment and the second data segment to obtain an updated reassembly context; determining a first size relationship between the first remaining space length of the updated reassembly context and a space length threshold; if the first size relationship indicates that the first remaining space length is greater than the space length threshold, repeatedly executing the splicing step until it is determined that the first remaining space length is less than or equal to the space length threshold to obtain a target reassembly context; and reassembling the data segments in the target reassembly context to obtain reassembled data segments.

The invention provides a multi-gateway collaborative mechanical data monitoring Internet of Things system and method for a sensor and a medium, and belongs to the technical field of data monitoring. The method comprises the following steps: collecting mechanical data through leaf nodes of a tree network, and converting the mechanical data into frequency signals; converging the frequency signals through routing nodes of the tree network, and segmenting a data packet containing the frequency signals into a plurality of data fragments with preset lengths; sending the data fragments to the sensing network platform through the sensing control platform; frequency signals are analyzed from the received data fragments, the frequency signals are converted into mechanical data, and the mechanical data are sent to the management platform; performing data recombination and integrity verification on the received mechanical data; and sending the complete mechanical data to a computer terminal, and controlling the computer terminal to render and generate an axial force and / or torque trend chart. According to the invention, accurate recombination and integrity verification of data after parallel transmission of multiple gateways can be realized, so that the reliability requirement of engineering safety monitoring is met.

The application discloses a full-network telemetry method based on telemetry metadata reorganization, and relates to the technical field of network communication.The full-network telemetry collects telemetry metadata of all network nodes in a specific range, and an INT collector constructs a data set by using INT telemetry reports of multiple continuous data packets; a telemetry metadata reorganization TMR algorithm is designed; telemetry metadata in the INT telemetry report is reorganized based on collected time and place, and TMR telemetry reports covering all network nodes are formed; a series of TMR telemetry reports obtained through the TMR algorithm processing are stored in a result set, and an accurate full-network view is formed.The application can more accurately reflect performance indexes of all nodes in the network, the performance is not affected by the length of a forwarding path, and is beneficial to network congestion identification and fault detection and the like.

The present application relates to the technical field of finance and banking, and discloses a supervision reporting integrated machine system based on software and hardware integration, which comprises a compiling module, an access module, a scheduling module and an aggregation module.The compiling module parses supervision specifications into a fixed-width compliance state bit string and binds the bit string to the physical core of a multi-core processor to establish a hardware verification topology path.The access module writes business flow data into a physical memory pool and allocates a data physical pointer and a compliance state bitmap.The scheduling module executes compliance verification logic by using the data physical pointer according to the hardware verification topology path, and updates the compliance state bitmap when the execution is passed.The aggregation module reorganizes business flow data into a huge physical memory page by using a data flow accelerator when the updated compliance state bitmap meets a mask condition.The sending module triggers an encryption accelerator card to generate ciphertext data based on the starting address of the huge physical memory page and sends the ciphertext data through a network interface card.The present application avoids cross-core copying of full data, reduces hardware bus communication delay, and improves the execution efficiency of concurrent verification.

The application discloses a method and system for improving neural network accelerator bandwidth optimization, and relates to the technical field of neural network accelerator, and the method comprises the following steps: performing convolution processing on input data through a convolution pipeline; performing activation function processing on the data processed through the convolution by using a single data processor; replacing traditional on-chip storage with an on-chip buffer for caching, and performing pooling processing and local response normalization processing on the data processed through the activation function by using a plane data processor and a channel data processor; and performing data format conversion on the data processed through the local response normalization by using a data reorganization engine. The application optimizes the data storage, reading and writing process of the accelerator during acceleration operation, effectively improves the utilization rate of the memory bandwidth, reduces the time consumed by single acceleration operation, and improves the performance of the accelerator.

The present application belongs to the field of aircraft telemetry communication technology, and particularly relates to a multi-aircraft PCM-FM telemetry system based on TDM-TDMA and a design method thereof. The system comprises an airborne synchronous transmitting unit, a ground single-station broadband receiving and processing unit and a ground parameter binding device. The airborne synchronous transmitting unit realizes unified time synchronization of space, air and ground through IRIG-B / Beidou dual-redundancy time service, transmits signals containing unique ID according to time slot, and is silent in the remaining period. The ground single-station broadband receiving and processing unit receives mixed signals and directly demodulates them by using single frequency point, single antenna and single channel, and multi-channel signal separation and TDM data reorganization are completed by recognizing ID of the upper computer to output standard IRIG PCM data stream. The implementation method comprises time synchronization, time slot allocation, preamble synchronization, time division transmission, receiving and demodulation, ID recognition, data reorganization and mode switching. The present application does not change the traditional system and ground station hardware, realizes single-station multi-target concurrent receiving, and solves the problems of traditional single receiving, low spectrum utilization and high transformation cost.

The application discloses a kind of quick man-machine interaction application systems and its use method and device, system includes IO module, DDM module and UI module, DDM module includes data analysis module, independent variable module, data reorganization module and data forwarding module, method includes: S1.IO module receives the data packet of external communication system and forwards data packet to DDM module and is parsed processing;S2.DDM module is parsed processing after the data packet from IO module, is forwarded to UI module;S3.UI module receives the data sent from DDM module after, carries out the visual display of data on the operation interface of UI module;S4.user views in UI module, and UI module sends the information data of operation to DDM module processing;S5.DDM module sends the information data handled to IO module;S6.IO module sends the data received to external communication system.The application solves the problem that user needs to learn complex programming language to realize man-machine interaction in prior art.

The invention provides a high-order mask comparison implementation method and system based on GPU acceleration, and relates to the technical field of post-quantum cryptography. Independently generating a random number for each thread at the GPU end; converting the input data from an arithmetic mask form to a Boolean mask form by using the random number; selecting a direct path or a bit slice path according to a data scale to carry out zero value detection: if the direct path is adopted, keeping an original data layout, multiplexing the same random number on a plurality of mask shares of the same data unit for processing, and judging whether the unit is zero or not according to an accumulation result; and if a bit slice path is adopted, reorganizing the data into bit slices, aggregating the same bits of a plurality of data units in each slice, performing differentiation processing and accumulation on the units in the slices according to the positions of the units, and realizing multi-unit parallel detection, thereby improving the performance and safety of high-order mask comparison in a GPU environment.

The invention relates to the technical field of laser radar signal processing, in particular to a high-speed signal collecting and processing system and method based on a laser radar, and the system comprises a high-speed data receiving module which receives and caches the original sampling data of the laser radar; the data recombination module is used for carrying out overlapping block recombination on the data; the parallel peak detection module is used for carrying out parallel peak detection according to a self-adaptive peak degree calculation method; the boundary arbitration module is used for carrying out conflict arbitration on the boundary candidate peak points; the dynamic threshold updating module is used for updating the dynamic threshold; and the assembly line scheduling module is used for carrying out multi-stage assembly line parallel scheduling. Through a data overlapping block parallel detection architecture, a self-adaptive peak degree calculation mechanism and pipeline parallel scheduling, high-speed real-time processing of laser radar signals is realized, the processing throughput and the environmental adaptability are improved, and the technical problems of low processing speed, serious boundary effect and low resource utilization rate in the prior art are solved.

The invention relates to a high-parallel three-dimensional array data recombination module and method supporting multiple operations. The data recombination module comprises an input processing module, a write address generation module, an RAM storage array, a read address generation module and an output processing module. And four three-dimensional array transformation operations of slicing, merging, compressing and transposing are realized. According to the module, a 512-bit wide data bus is adopted to be matched with 32 parallel RAM channels, and assembly line processing is achieved in combination with two ping-pong RAM sets. By means of rearrangement, displacement and address generation logic which are finely designed, efficient data reading and writing of 32-byte alignment are kept, and meanwhile flexible dimension transformation with one byte as the minimum unit is supported. The problems of low conversion efficiency and high power consumption of a traditional CPU are effectively solved, data conversion delay and power consumption are reduced by dozens of times, and an efficient data recombination solution is provided for high-performance computing application such as a three-dimensional convolutional neural network.