69results about How to "Reduce time overhead" patented technology

The invention relates to an optimal path selection method for data exchange in a cross-network scene, and belongs to the field of computers. The method comprises the following steps of: initializing a communication relationship among nodes in all network systems exchanged and associated across the network systems into a data exchange routing information table; enumerating a switching path and calculating a path length; obtaining QoS attribute vectors of the enumerated exchangeable path link, including response time, reliability and stability, as link quality parameters, and meanwhile, comprehensively forming an exchange path attribute parameter by taking an exchange path length as a security evaluation factor; weight is distributed to each parameter, and optimal path calculation is carried out, so that the optimal path of data exchange in the cross-network scene is obtained. According to the method and the device, the data exchange can be completed when the direct path is blocked in the cross-network scene without opening the exchange link, so that the economic cost caused by opening the link and the time overhead of offline exchange can be effectively saved.

The invention provides an operator execution method and device, a storage medium and a program product, and relates to the technical field of artificial intelligence, in the method, when a second thread bundle group adopts a vector operator to carry out vector calculation on an nth batch of tasks, a first thread bundle group adopts a first matrix operator to carry out matrix calculation on an (n + 1) th batch of tasks, and a second thread bundle group adopts a second matrix operator to carry out matrix calculation on the (n + 1) th batch of tasks; meanwhile, the third thread beam group adopts a second matrix operator to perform matrix calculation on the (n-1) th batch of tasks, and the first thread beam group and the third thread beam group are both composed of tensor calculation units, so that the tensor calculation units can process two batches of tasks at the same time. The tensor calculation unit and the vector calculation unit can process different batches of tasks at the same time, so that the mutual waiting time is avoided, and the time overhead caused by processing multiple batches of tasks by the Attention operator can be reduced.

The invention discloses a scene depth estimation method and device, image processing equipment and a storage medium, and belongs to the technical field of computer vision, and the scene depth estimation method comprises the steps: selecting a single image with shielding, reflection or perspective from a multi-view image set for depth fusion; the improved depth information is spread to other views in a re-projection mode, so that the time overhead is reduced while the reconstruction effect is improved; performing semantic division on the index image to obtain the weight of each semantic factor; the problem that feature points are difficult to match due to a special area is solved by introducing single-view depth estimation; fusing a single-view estimation result and a multi-view depth estimation result under the constraint of semantic information, and spreading the fused result to other view angles in a re-projection manner to obtain a consistent depth estimation result corresponding to the current scene; consistency estimation can be accurately carried out on the depth of each area in the scene.

The invention relates to an IP and IPFC bridging device based on an FPGA, belongs to the technical field of Ethernet network and FC-AE network communication, and solves the problem of complexity of Ethernet network and FC-AE network communication equipment in the prior art. The bridging device comprises an IP data transmission module used for analyzing an IP data frame sent by an Ethernet network, transmitting the IP data frame to a conversion control module, and transmitting the IP data frame converted by the conversion control module to the Ethernet network; the IPFC data transmission module verifies the FC data frames sent by the conversion control module, transmits the FC data frames without errors to the FC-AE network, and transmits the FC data frames sent by the FC-AE network to the conversion control module; the conversion control module converts effective data sent by the IP data transmission module into FC data frames, sends the FC data frames to the IPFC data transmission module, converts the FC data frames sent by the IPFC data transmission module into IP data frames, and sends the IP data frames to the IP data transmission module. Communication between the Ethernet network and the FC-AE network is realized, and the device is simple.

The invention provides a power consumption data hierarchical federated learning method, device and equipment and a storage medium, and the method comprises the steps: enabling a participant to obtain an initial model parameter, and starting the training based on the initial model parameter; the household energy management system collects real-time power consumption data, performs local training based on the real-time power consumption data, updates local model parameters of the household energy management system, and submits the local model parameters to the load integrator of the corresponding area; and the load integrator performs asynchronous local aggregation based on the local model parameter and the current global model parameter to obtain a local aggregation parameter, and issues the local aggregation parameter to the home energy management system of the corresponding region. According to the invention, the number of lazy clients is reduced, the overall efficiency of federal learning is improved, and the problem of poor enthusiasm of participants in the prior art is solved.

The embodiment of the invention provides a virtualization manager upgrading method and device and a storage medium. According to the embodiment of the invention, in the upgrading process of the virtualization manager, when the state of the virtual instance is stored and recovered, the memory address of the state information of the virtual instance is transmitted between the original virtualization manager and the upgraded new virtualization manager; compared with the traditional scheme that the virtual instance state information needs to be copied back and forth between the virtualization management process of the user mode and the virtualization manager of the kernel mode, the method has the advantages that the copied data volume between the original virtualization manager and the new virtualization manager is reduced, and the time overhead in the state recovery process of the virtual instance is reduced; and the pause time of the virtual instance generated by upgrading the virtualization manager can be reduced.

The invention discloses a MySQL database migration method and system based on a block file, and the method comprises the steps: enabling a snapshot file generation service to generate a snapshot for a source database directory, storing the snapshot in a designated local disk directory, and recording a timestamp generated by the current snapshot; the snapshot file transmission service transmits the snapshot to a target library server; the incremental data extraction service obtains a snapshot generation timestamp and analyzes an incremental log after the timestamp, binary data in the incremental log is converted into structured data, and the structured data is stored in a local disk directory; the incremental data conversion service is used for converting incremental structured data into standard DML statements of a database by identifying data operation content and analyzing to adapt to all field types; and reading and sequentially executing the DML statements through the incremental data writing service, and restoring the data of the source database to the target database. On the basis of block file migration, the thermal migration of the database is realized in combination with incremental logs.

The invention belongs to the technical field of computer software and data processing, and provides a differential upgrading method for a compressed file, which is divided into a differential stage and a reduction stage: the differential stage: decompressing an original compressed package of a starting version and an original compressed package of a target version; unifying and standardizing file permissions for the two decompressed folders, and modifying, accessing and changing timestamps; re-compressing to generate a first new compressed file and a second new compressed file with byte-level certainty; generating a first difference file based on the difference between the standardized directory structure and the content; comparing the second new compressed file with the original compressed file of the target version, performing lossless analysis on the ZIP metadata, and performing structured modeling on difference information to generate a second difference file; in the reduction stage, standardization, first differential reduction, recompression and second differential driven metadata accurate reconstruction are sequentially executed, and a result completely consistent with the byte level of the original compressed file of the target version is output. According to the method, the differential compression ratio and the restoration fidelity can be improved.

The invention relates to a cluster scheduling method and device, computer equipment, a computer readable storage medium and a computer program product, and relates to the technical field of cloud native. According to the invention, under the condition of not changing the structure of the cluster and the scheduling system thereof, the influence of the scale expansion of the cluster on the cluster scheduling performance can be effectively dealt with. The method comprises the following steps: acquiring a target job, finding out a target node meeting a running condition of the target job in a plurality of partitions according to a partition priority sequence and node states and node snapshot data of nodes, and scheduling the target job to the target node, and updating node snapshot information of the target node in the node snapshot data and node real-time data of the target node in the node real-time data according to a scheduling result of the target node.

The invention provides an operating method of a memory device and a memory device thereof. The memory device can be a three-dimensional NAND flash memory with high capacity and high performance. The operating method includes performing a program operation on the plurality of memory cells based on the i-th potential state. The programming operation based on the ith potential state comprises the following steps: in a programming pulse stage, applying a programming pulse corresponding to the ith potential state to a plurality of first storage units corresponding to the ith potential state; and, in a verification stage after the program pulse stage, performing a first sensing operation on the plurality of second memory cells in the (i-1)-th potential state to verify a threshold voltage distribution situation of the plurality of second memory cells, and performing a second sensing operation on the plurality of first memory cells to implement a program verification operation in the i-th potential state.

The invention discloses a grid exhaustion search cross-pulse pairing method based on correlation matching, and belongs to the technical field of radio signal positioning and detection. According to the method, aiming at the problems of large cross-pulse signal pairing calculation amount and low accuracy in a multi-aircraft cooperative positioning system, efficient and accurate cross-pulse pairing is realized by establishing a cross-pulse boundary model, a symbol matching constraint mechanism and a related matching verification mechanism. The method comprises: determining upper and lower bounds of a cross-pulse according to a geometrical relationship; correcting the search range by using symbol matching; solving a candidate target solution in the limited interval by adopting grid exhaustion search; and a unique effective solution is screened through time and geometry consistency verification. The method can effectively reduce the calculation complexity, improves the cross-pulse pairing precision and uniqueness, and is suitable for a passive positioning system in a multi-target, multi-base-station and low-signal-to-noise-ratio environment. Experiments prove that the pairing time can be shortened by about 50% while high positioning precision is kept, and the method has high engineering application value.

The invention discloses a three-dimensional terrain reconstruction method based on an elevation image, electronic equipment and a storage medium. The method comprises the following steps: (1) loading data; (2) edge detection; (3) concurrent mapping; and (4) performing three-dimensional reconstruction. According to the method, the edge detection technology is adopted, different areas on the topographic map are divided, adaptive down-sampling is carried out according to the categories of the different areas, and the elevation values of the sampling points are converted into height information, so that the elevation image of the terrain is reconstructed into a visual three-dimensional image; on the premise that the visualization effect is maintained to be equivalent to that of the existing elevation image, a large amount of redundant data is reduced, so that the time overhead is greatly reduced.

Aiming to solve the technical problem that the time overhead of mobile entity trajectory data fusion processing increases significantly under the condition of a sharp increase in data traffic, which may cause system crash, the present invention provides an efficient trajectory fusion method based on local matching, which includes the following steps: initial parameter setting, trajectory data rasterization cache processing, cache data life detection, trajectory data sequence matching calculation; matching success condition determination. The present invention can significantly reduce the number of trajectory similarity calculations by pre-grouping mobile entity trajectory data, thereby improving the efficiency of fusion processing. It is suitable for scenarios with diverse trajectory data sources, large data traffic, and wide geographical distribution of trajectories.

The invention relates to an interrupt system verification method and device and a storage medium. The method comprises the following steps: acquiring a first standard file, wherein the first standard file comprises interrupt signal information; analyzing based on the first standard file to obtain a second standard file matched with the verification tool in the preset form, the second standard file comprising transmission path description of the interrupt signal; and performing interruption path correctness verification based on the second standard file by using a preset form verification tool to obtain an interruption path verification result. According to the embodiment of the invention, the method facilitates the maintenance and modification of the interrupt system, guarantees the iteration quality, enables different interrupt system verification items to be reused, can achieve the automatic interrupt system verification process, and improves the verification efficiency. The time overhead and machine resource overhead (including machine computing resources and machine storage resources) of interrupt system verification are greatly reduced, the workload is reduced, and the error probability is reduced.

The invention discloses an index-based content awareness influence maximization query response method and device. The method comprises the following steps: 1) establishing an index node selection mechanism based on minimum vertex coverage approximation; 2) constructing a space-efficient index table to store content perception reachability information; 3) designing a reachability table estimation method to calculate a reachability probability lower bound between node pairs; and 4) providing pruning accessibility table estimation and an inertia forward greedy algorithm to optimize query efficiency. Verification on Wiki, Twitter and other real network data sets shows that the propagation effect of the method is similar to that of an existing optimal influence maximization method, and the query response speed is increased by 1-2 orders of magnitude.

The present application provides a method, device and equipment for obtaining a question-answer pair, the method comprising: obtaining the row title, column title and cell content corresponding to the candidate cell; generating a first question and a second question based on the row title and the column title, wherein the row title in the first question is located in front of the column title, and the column title in the second question is located in front of the row title; determining a reference question based on the semantic coherence of the first question and the semantic coherence of the second question, if the semantic coherence of the first question is greater than that of the second question, the reference question is the first question; if the semantic coherence of the first question is less than that of the second question, the reference question is the second question; generating an extended question-answer pair based on the reference question and the cell content, wherein the question of the extended question-answer pair is the reference question, and the answer to the extended question-answer pair is the cell content; generating a target question-answer pair based on the extended question-answer pair. Through the scheme of the present application, the diversity of question-answer pairs is ensured, and the generation quality of question-answer pairs is improved.

This invention relates to an FPGA-based method for correcting endoscopic images, an endoscopic system, and a medium. The FPGA-based method for correcting endoscopic images includes: acquiring the current spatial attitude data of the endoscopic probe using an FPGA chip; calculating the current roll angle of the endoscopic probe based on the spatial attitude data; calculating the position coordinates of a target pixel in the original image based on the roll angle and the pixel coordinates of the target image; calculating the pixel value of the target pixel based on the position coordinates of the target pixel in the original image and the pixel values ​​of at least four pixels in the pre-cached original image; and integrating the pixel coordinates of the target image and the pixel value of the target pixel to obtain the target image.

This application discloses an image file management method, including: receiving a write instruction used to instruct to write a first image file; determining, based on the write instruction, at least one first edge cloud node from a plurality of edge cloud nodes included in an image file management system; and writing a plurality of data blocks included in the first image file into the at least one first edge cloud node, so that each first edge cloud node stores a part or all of the plurality of data blocks included in the first image file. In this application, time overheads for downloading an image file are reduced, and a download speed of the image file is improved.

The invention belongs to the field of scientific satellite data processing, and discloses a scientific satellite data processing system of a cross-APID hybrid dependency relationship, which comprises an original data preprocessing module used for preprocessing multi-source data and generating a 0-level product; the engineering parameter template and dependency matrix analysis module is used for analyzing the 0-level product according to the scientific satellite task engineering parameter template of the parameter analysis configuration file and the parameter dependency matrix of the cross-APID source packet; the Redis cache library sharing architecture module is used for sharing each data product parameter value of the cross-APID source packet by using Redis cache; the cross-APID dependency parameter scheduling module is used for scheduling a processing sequence according to different data types and dependency conditions; and the data product formatting generation module is used for generating various types and levels of products based on the product definition specifications after all the parameters are processed and cached.

The embodiment of the invention discloses a memory management method and device, equipment and a computer storage medium, and the method comprises the steps that network equipment allocates a first memory space corresponding to a first window threshold value for a radio link control layer (RLC, Radio Link Control) entity in an acknowledgement mode; the first window threshold value is smaller than or equal to the size of a sending window of the RLC entity; when the use condition of the first memory space meets a first condition, allocating a second memory space; the first memory space and the second memory space are both used for caching related data of an RLC entity sent to a terminal by the network equipment.

The application provides a dynamic data chart statistical method and device, and relates to the technical field of big data. The method comprises the following steps: receiving a data query request of a user; querying full data corresponding to the data query request from a server backend according to the data query request; storing the full data in a data visualization chart library corresponding to a pull-down box selector of a front-end interface; the pull-down box selector provides a plurality of statistical index options of user-specified data in the full data for the user; after the user selects a first statistical index option from the pull-down box selector, extracting the user-specified data from the data visualization chart library, inputting the user-specified data and the first statistical index option into a data statistical analysis model, and outputting a first statistical analysis result; visually rendering the first statistical analysis result to generate a first visualization chart; and showing the first visualization chart to the user.

An MIMO array sparse rapid design method of a fuzzy function generator based on deep learning comprises the following steps: firstly, directly learning a fuzzy function mapping relation corresponding to virtual array configuration and an incident angle by using a deep learning model so as to replace a traditional analysis fuzzy function calculation process depending on a manifold matrix; and then performing efficient performance evaluation on the candidate sparse array based on the fuzzy function generator, and obtaining a sparse array structure meeting the requirements of high resolution and low sidelobe in combination with an optimization strategy. By means of the technology, acceleration of more than 35 times can be achieved on the premise of not explicitly calculating the fuzzy function, the fuzzy function prediction root mean square error (RMSE) lower than 0.1 is kept, and therefore the sparse array design time is greatly shortened. Experiments on a 2T4R K-band MIMO radar prove that the designed array achieves 10 dB sidelobe level in a 100-degree field of view, the angular resolution is improved by about 32.8%, and the effectiveness and engineering application value of the method in high-resolution MIMO sparse array design are proved.

Embodiments of the present application provide a data classification method and device, processing equipment and storage medium, relating to the field of security science and technology. The fingerprint of the data to be classified is segmented to obtain a plurality of to-be-matched fingerprint segments; a candidate fingerprint set is determined according to the results of the segmented search of each to-be-matched fingerprint segment in a preset database, the candidate fingerprint set including a plurality of candidate fingerprints; each candidate fingerprint in the candidate fingerprint set is matched with the fingerprint of the data to be classified to determine a target fingerprint, and the classification result of the data to be classified is determined according to the target fingerprint. The fingerprint of the data to be classified is segmented and then searched in the preset database to determine the target fingerprint, which, compared with the way of searching and matching the whole fingerprint in the prior art, reduces the search and matching overhead of the system, reduces the search time and improves the efficiency.

This invention provides a multi-channel continuous fast packet reception, sorting, and lock-free asynchronous read / write method, relating to the fields of computer communication and storage technology. The method includes: employing a dynamic expansion mechanism to continuously receive and store multi-channel data packets; classifying and sorting the stored data packets; and asynchronously reading and writing the parsed data packets in a lock-free manner. This invention generates an independent message stack for each channel, enabling continuous reception and storage of messages from multiple channels without pre-allocating excessive memory space. The speed of receiving buffer expansion is dynamically adjusted based on data flow rate. Fast classification and sorting of data packets is achieved using a hash table or sorted binary tree. A circular stack is created to sequentially store the parsed data content of each packet, updating the stack top position after data content is stored. When reading data, the starting position for reading is found using the stack top position and the number of packets read, and then the data is copied, thereby achieving lock-free asynchronous parallel data read / write.

The invention discloses a fast Paillier encryption method in a federated learning system, comprising: step 1, offline obtaining a secret state zero set by modular exponentiation operation: in the offline stage of the training protocol, each participant of the federated learning system determines the total number of its own encryption tasks according to its corresponding training tasks, and each participant locally uses a public key to generate a number of secret state zeros corresponding to the total number of its own encryption tasks by modular exponentiation operation to form a secret state zero set; step 2, using the secret state zero set obtained offline to encrypt online by modular multiplication operation: in the online stage of the training protocol, each time each participant performs Paillier encryption on plaintext, a secret state zero is randomly selected from the secret state zero set generated by the participant to perform this Paillier encryption operation by modular multiplication operation, and the selected secret state zero is removed from the secret state zero set. The method can reduce the time overhead of encryption operation in the online stage, and effectively improve the operation efficiency of the online stage of the federated learning training protocol.

The present invention relates to a method for extracting and eliminating incomplete routes based on ADS-B data, which belongs to the field of track processing technology, and includes the following steps: step 1, pre-processing regional sample data to generate partially completed routes; step 2, for track segments that cannot be classified as a complete cluster in step 1, iteratively fusing similar tracks to generate routes; step 3, forming a regional route set by combining the routes in step 1 and the routes in step 2; step 4, extracting the latitude and longitude position information of unidentified target track points, matching the nearest routes with the regional route set generated in step 3, calculating the position similarity between the unidentified target track and its nearest routes, and determining whether the target is flying on a civil aviation route; step 5, screening the targets on the civil aviation route determined in step 4 for sensitive targets. A new perspective for the sensitive target screening method is provided, with high accuracy and stable performance in complex scenarios.

The present disclosure relates to a method, device and storage medium for determining a reference signal of a candidate cell. The method includes: a network device sends a first indication information, and the first indication information is used to indicate the reference signal of the candidate cell that the terminal device needs to perform beam measurement. In this way, the terminal device can directly perform beam measurement on the reference signal of the corresponding candidate cell indicated by the network device, thereby solving the problems of large measurement time overhead and high complexity caused by the terminal device performing beam measurement on the reference signals of all candidate cells configured in the prior art.

This disclosure relates to a simulation method, electronic device, and computer-readable storage medium for semiconductor devices. The method includes: constructing a multi-branch tree structure based on meshed data of a meshed semiconductor device, the multi-branch tree structure including at least a root node, multiple layers, and multiple leaf nodes, each leaf node representing a corresponding polyhedron among multiple polyhedra in the meshed semiconductor device; in response to determining a first spatial position of a particle within the semiconductor device, determining a first search path from the root node to a first leaf node among the multiple leaf nodes based on the correspondence between each leaf node and its corresponding polyhedron; and in response to the particle moving from the first spatial position to a second spatial position within the semiconductor device, determining a second search path in the multi-branch tree structure based on the first search path to determine a second leaf node representing the polyhedron where the second spatial position is located. This method effectively improves the performance of ion implantation simulation.

The embodiment of the present application provides an operator execution method, device, storage medium and program product, which relates to the field of artificial intelligence technology. In the method, an artificial intelligence chip is used to use a multi-head attention mechanism operator to execute M1 tasks, and the multi-head attention mechanism operator includes N attention mechanism operators; the first processing unit determines K1 subtasks to be executed, and any of the K1 subtasks to be executed is a partial task of executing one of the M1 tasks using the attention mechanism operator, and the first processing unit uses the attention mechanism operator to execute the K1 subtasks to be executed respectively. In the above process, multiple subtasks to be executed belonging to the same task can be assigned to different processing units, and different processing units use the attention mechanism operator to execute the subtasks to be executed in parallel. Therefore, the time overhead caused by the processing unit using the multi-head attention mechanism operator to execute the task can be effectively reduced.

The invention provides a 3D point cloud data processing method and device based on a mixed KD tree and related equipment, and aims to solve the problem of low search processing efficiency of 3D point cloud data in the prior art. Judging whether a preset historical query set contains historical query point cloud features which are the same as or similar to the reference point cloud features, and if the preset historical query set does not contain the historical query point cloud features which are the same as or similar to the reference point cloud features, according to preset leaf nodes which are contained in a preset mixed KD tree and are based on a disordered set, determining whether the preset historical query set contains the historical query point cloud features which are the same as or similar to the reference point cloud features; the preset leaf nodes are searched in parallel by adopting the preset K nearest neighbor algorithm to obtain the target point cloud feature corresponding to the reference point cloud feature, so that approximate search based on historical query and accurate search based on the mixed KD tree are combined, parallel search of the mixed KD tree is realized, the search processing efficiency of the 3D point cloud data is improved, and the search efficiency of the 3D point cloud data is improved. The method can be applied to financial science.