58results about How to "Reduce time complexity" patented technology

The invention relates to the technical field of wiring, and discloses a wiring method, electronic equipment and a computer readable storage medium, and the method comprises the steps: converting an initial layout layout of a circuit board into a grid chart; according to the grid chart, a coarse wiring result of the initial layout layout is determined, and the coarse wiring result is an initial wiring path meeting a wiring task and achieving a wiring cross minimization target; generating a wiring geometric figure according to the safety spacing and the initial wiring path, and generating a device geometric figure according to the safety spacing and component information in the initial layout layout; and under the limitation of the geometric figure of the device, optimizing the position of the geometric figure of the wiring by taking the wiring constraint as a target to obtain a fine wiring result. According to the method, the wiring process is divided into a coarse wiring stage meeting the goals of connectivity and minimum wiring intersection number and a fine wiring stage meeting the wiring constraint, so that the solving complexity can be reduced, and a wiring scheme with the minimum wiring intersection number can be obtained in a relatively short time.

This invention belongs to the technical field of embedded artificial intelligence and real-time system scheduling, and more specifically, relates to an adaptive load resource scheduling method and apparatus for embedded GPUs. The method involves offline acquisition and modeling of the platform's thermal throttling behavior, constructing DNN variants with different computational loads and accuracies, and profiling them under multiple frequency and TPC configurations to obtain a latency-power mapping table. In the online phase, based on task deadline margins and current temperature / frequency states, the method looks up the variant with the highest accuracy and meeting the deadline under the current thermal conditions and selects the corresponding TPC configuration from the profiling table. If necessary, phased online pruning and low-overhead switching are implemented, and the method collaborates with system-level DVFS / task mapping to optimize energy efficiency. This method can still guarantee the real-time performance of inference tasks under passive frequency reduction conditions and significantly reduce performance fluctuations caused by thermal throttling.

The invention discloses a device upgrading method and a device upgrading system. The device upgrading method comprises the following steps of: establishing a network connection between the control module of the display device and the information management function module of the remote server through the switch module of the display device; the device information is transmitted to the information management function module through the control module; performing verification processing on the device information through the information management function module, and generating device upgrading information to the control module; the control module downloads the device upgrade data from the information management function module according to the device upgrade information, verifies the device upgrade data, and upgrades the function device of the display device according to the device upgrade data. According to the device upgrading method and the device upgrading system, the device upgrading efficiency can be effectively improved, the time cost and complexity of device upgrading are reduced, the use experience is improved, the process of device upgrading and maintenance is perfected, and the use convenience is improved.

The application discloses an active intrusion prevention method, relates to the technical field of system network security, and aims to solve the problems of low detection precision and detection lag of existing network abnormal intrusion, and comprises the following steps: obtaining preprocessed network data; performing multi-dimensional traffic data analysis on the network data to obtain the network data labeled with network state credibility measurement results and various sub-dimension measurement labels; calculating the labeled network data through a machine learning classification algorithm to obtain characteristic behavior classification of the network data; and performing suspicious node detection through fine-grained real-time measurement according to the classification results and updating a security strategy. The application further discloses an active intrusion prevention device. The application can quickly analyze network intrusion through multi-dimensional analysis of network data and machine learning.

A neural network graph partitioning system and method are disclosed. The graph partitioning system is used to partition a neural network graph into a series of subgraphs and allows multiple subgraphs to be executed in available hardware subsystems. The system estimates the computation time and memory bandwidth of the partitioned subgraphs based on a cost function. The graph partitioning system is a hardware cycle estimation model that can be run quickly and implements parameterization of memory latency. The graph partitioning system supports heterogeneous partitioning for different types of accelerators, such as central processing units, graphics processing units, application specific integrated circuits, etc. A method of partitioning a neural network graph into a series of subgraphs is also disclosed.

This invention discloses a text summarization method and system based on sparse attention acceleration, comprising: reading long text data, performing word segmentation and embedding encoding processing, extracting sequence feature vectors, and mapping them to a query matrix Q, a key matrix K, and a value matrix V; constructing a summarization generation network, including a sparse attention calculation module, a feedforward calculation module, a prediction head module, and a key-value caching module; inputting the query matrix Q, key matrix K, and value matrix V into the summarization generation network, first passing through a backbone network composed of multiple stacks of the sparse attention calculation module and the feedforward calculation module, and then being processed by the prediction head module, while the key-value caching module caches historical decoding states in real time to obtain an initial text feature vector; based on the initial text feature vector, performing an autoregressive decoding process through the summarization generation network to output the final summary result. This invention can accelerate the decoding process and ensure the semantic integrity and coherence of the generated summary.

The application relates to the technical field of image processing, and discloses an image processing method and device based on a double-resolution decomposition tree, and a medium. The method comprises the following steps: acquiring computer vision image data; obtaining a generated tree structure according to the computer vision image data; processing the generated tree to obtain a dual solution and constructing an adjustment cost matrix; when it is judged that the cost matrix does not meet the requirements, performing an Euler circuit serialization mapping operation to obtain a maintenance cost matrix; dividing the matrix through a double-resolution matrix decomposition tree to obtain a plurality of small-scale blocks and medium-scale blocks; at the same time of the division, lazy labels are also lowered into the medium-scale blocks and the small-scale blocks; dividing the small-scale blocks with labels into constant blocks, and finally merging the constant blocks to obtain an optimized matrix; using an optimal transport algorithm to process the computer vision image data, and using the optimized matrix to reduce the time complexity of single iteration maintenance of the optimal transport algorithm.

The application discloses a kind of relics fragment similarity measurement methods based on multi-feature information, including collection relics fragment data, clustering relics fragment image, select a kind of clustering after relics fragment image conversion into binary image, extract the contour line of binary image one by one, identify the corner point of contour line and identify the similarity of two relics fragment image step.The application is according to the color information and texture features of ceramic fragments, before fragment splicing, different characteristic ceramic fragments are clustered, then the edge superpixel features and contour curvature features of ceramic fragments are fused, the similarity between fragments is calculated using difference and distance measurement formula, effectively reduce the time and complexity of fragment splicing, improve the accuracy of fragment similarity comparison.

The utility model relates to the technical field of connectors, and discloses a sealing structure for an assembled high-speed connector, which comprises a first connecting pipe and a second connecting pipe, the outer part of the first connecting pipe is provided with raised lines, the inner part of the second connecting pipe is provided with a connecting assembly for connecting the first connecting pipe and the second connecting pipe, and the connecting assembly is provided with a sealing structure for sealing the first connecting pipe and the second connecting pipe. And a cavity is formed in the first connecting pipe, a plurality of dampers are fixedly connected to the interior of the cavity, and springs are arranged on the exteriors of the dampers in a sleeving mode. According to the utility model, the sealing ring is in contact with the inner wall of the second connecting pipe, so that the interior of the second connecting pipe can be sealed, and when the sealing ring is in contact with the inner wall of the second connecting pipe, the inner wall of the second connecting pipe can press the sealing ring, so that the sealing ring can press the damper and the spring; and the sealing ring can be attached to the inner wall of the second connecting pipe more tightly under the reset acting force of the damper and the spring, and then the sealing effect of the sealing ring can be improved.

The application provides a machine tool abnormality monitoring method based on context double-layer clustering and incremental learning, and relates to the technical field of intelligent manufacturing and equipment health monitoring.The application introduces the concept of processing context, establishes multiple micro-working condition clusters based on the multi-source time sequence data of the processing history parts, and constructs the health baseline model of each micro-working condition cluster; based on the time sequence feature sequence of the current processing process, the target micro-working condition cluster of the current processing process is determined, and whether the current processing process of the numerical control machine tool exists abnormality is judged based on the health baseline model of the target micro-working condition cluster; when an unknown new working condition appears, the micro-working condition cluster and the corresponding health baseline model of the unknown new working condition are established; when the feature gradual drift of any one micro-working condition cluster appears, the health baseline model of the micro-working condition cluster is updated.When the numerical control machine tool processes a new part after changing production, a new working condition monitoring baseline can be generated without relying on manual labeling data and full amount retraining, which greatly enhances the flexible adaptation ability of the numerical control machine tool.

The invention discloses a medical insurance cross-hospital prescription compliance detection method and system based on a date counting slot. The method comprises the following steps: acquiring prescription information of a current prescription; obtaining a cross-hospital historical drug preparation record, corresponding to the general drug name, of the insured in a preset time window; determining a corresponding medication time interval according to the historical medication date and the historical medication day number for each historical medication record, and performing counting operation on a preset date counting slot based on the medicine return identifier; traversing a date counting slot, and determining a first effective medicine opening date corresponding to the general name of the medicine by accumulating a counting value; and judging whether the current prescription forms illegal advanced medicine preparation or not based on the first effective medicine preparation date, the current medicine preparation date and a preset threshold value of the number of days for allowing advanced medicine preparation. Unified marking and accumulation calculation are carried out on the intervals through the date counting slot, linear counting and accumulation based on a fixed time window are achieved, the time complexity of an algorithm is reduced, and efficient real-time judgment of illegal advanced medicine preparation is achieved.

The application provides a machine translation model generation method and device, comprising: obtaining training data; training a machine translation model based on a pre-stored deep neural network using the training data; the machine translation model is used to translate a text to be translated into a text translation; the machine translation model comprises a memory enhancement adapter layer, a first memory and a second memory; the memory enhancement adapter layer is used to retrieve information in the first memory and / or the second memory and utilize the information; the first memory is used to save source language data obtained by performing reverse translation and forward calculation processing on the training data; and the second memory is used to save target language data obtained by performing reverse translation and forward calculation processing on the training data. The application saves the training data in a form acceptable to the model, reads information from the training data to assist the model when performing a translation task, and achieves the effect of adapting to various translation scenarios while taking into account lower algorithm time complexity and better model performance.

The application discloses a beamforming method of a holographic MIMO antenna array based on a steering vector, and under the condition that a user end position is unknown, multiple code words are constructed based on the steering vector through the combination of the pitch angle and the azimuth angle of the user end relative to the antenna array, a code word with the maximum gain of a receiving end is selected from the code words according to the feedback of the user end, and the code word is used as the optimal transmitting or receiving coefficient of the antenna array. The user end only needs to feed back the corresponding information in sequence, the requirement for the feedback time delay is slightly low, the error code of the feedback information has a certain fault tolerance, the time complexity is relatively low compared with the exhaustive search method, the performance is good, the outdoor channel changing rapidly can be well responded, the robustness to the channel fluctuation is higher, and the overall power gain is improved. Moreover, under the condition that the user end position is known, the transmitting or receiving coefficient of the antenna array can be directly determined.

The application discloses a kind of cloud data center high efficiency container arrangement method and system, it includes obtaining cloud data center overall information, including server quantity, each server resource upper limit, each server minimum power, each server maximum power, average migration network bandwidth budget provided by cloud data center;The information of all containers in a continuous time slot is obtained, including container quantity, each container resource request amount, each container image size, each container running time, each container migration attribute;According to the information of all containers in the continuous time slot and cloud data center overall information, the optimization problem with the minimum cloud data center average energy consumption as the target is constructed and solved, and the container arrangement decision of the continuous time slot is obtained.The application can obtain current time slot container arrangement decision under the condition of meeting various resource constraint conditions in the case where future container creation request quantity and container specification parameters cannot be accurately obtained, and is efficient, energy-saving and environment-friendly.

This application belongs to the fields of artificial intelligence and fintech, and relates to an AI-based customer classification method, including: acquiring customer data and constructing a topological relationship network corresponding to the customer data; inputting the topological relationship network into a graph convolutional neural network for training to obtain a trained graph convolutional neural network; acquiring a target topological relationship network corresponding to target customer data, and processing the target topological relationship network through a customer segmentation model to generate segmentation results. This application also provides an AI-based customer classification device, computer equipment, and storage medium. Furthermore, this application relates to blockchain technology, and the customer segmentation model can be stored in the blockchain. The customer classification method of this application can be applied to user identification in the financial field, requires no labeled data for training, has low time complexity, and the use of the customer segmentation model can achieve fast and accurate generation of customer segmentation results.

The application discloses a neural network-based electromagnetic field simulation grid self-adaptive generation method. The application performs grid partitioning on a semiconductor device simulation model, obtains original partitioning data, and constructs node features; a non-directed graph is constructed, the graph nodes of which adopt the node features of the grid points, and the connecting edges adopt the Euclidean distances between the grid points and adjacent grid points; the non-directed graph is input into a double-branch neural network, and a predicted node position is output; each node of the original partitioning data is corrected by using the predicted node position obtained by the double-branch neural network, and new partitioning data is formed to perform electromagnetic field simulation. The method can not only efficiently adjust the node density by fusing physical field information, but also strictly maintain the geometric boundary and topological structure characteristics of the device.

The application discloses an ICS intrusion detection system and method fusing reinforcement learning and feature selection optimization, carries out binary coding and population initialization on data feature selection of an industrial control system (ICS) historical data set, carries out offline training through SVM-reinforcement learning, takes the accuracy obtained on a verification set as a fitness function, designs cross operation and mutation operation based on cumulative probability to update the population, and obtains an optimal feature set after iterative optimization; feature selection is carried out on an ICS real-time data set based on the optimal feature set, online intrusion detection testing is carried out on the real-time data set through support vector machine (SVM)-reinforcement learning, and thus an intrusion detection performance index is obtained. The application adopts a new mode based on SVM-reinforcement learning, and fuses intelligent optimization of optimal feature selection on this basis, and improves the intelligent design level and the precision of intrusion detection of the ICS intrusion detection system.

This invention discloses a signal detection method for an OTFS system based on an adaptive wavelet scattering convolutional neural network, comprising: converting the data symbols to be transmitted to the DD domain, and then obtaining an OTFS signal through an OTFS modulator; the OTFS signal reaching the receiver through a wireless channel; at the receiver, transforming the received signal back to the DD domain using an OTFS demodulator; extracting the real and imaginary signal matrices from the received OTFS signal, performing a three-layer wavelet scattering network transformation on these two matrices to obtain scattering feature coefficients for each layer, concatenating the scattering feature coefficients along the feature channel direction to reconstruct a high-dimensional feature matrix of the signal, and using it as the input to an adaptive wavelet convolutional neural network 2D-AWCNN to obtain the possible values ​​corresponding to each QAM symbol; and performing 16-QAM demodulation on the output of the adaptive wavelet convolutional neural network 2D-AWCNN using a demodulator to obtain the output bits used for signal detection.

The application discloses an airborne Lidar point cloud building extraction method and system based on space profile sampling guidance, a medium and equipment, and belongs to the technical field of three-dimensional point cloud segmentation. Airborne radar point clouds are acquired, a segmentation network model is trained, three-dimensional point cloud segmentation is performed based on the trained segmentation network model, and buildings are extracted. In the encoding layer, random sampling and feature aggregation are performed, the point cloud scale is compressed, and the point cloud features are enhanced. In the feature aggregation, the nearest neighbor points are searched based on the KNN algorithm and the space profile algorithm respectively, then, spatial position coding and feature coding based on the self-attention mechanism are performed in sequence, finally, the enhanced features are obtained through attention pooling, and the point cloud scale is compressed to 1 / 256 of the original. In the decoding layer, the compressed point cloud is up-sampled and jump-connected using the nearest neighbor interpolation method, the point cloud scale is restored, and finally each point is classified. The application can significantly improve the classification accuracy of buildings in large-scale airborne point clouds, has a low time complexity, and has high classification accuracy and efficiency.

This application relates to a map construction method. Addressing the technical problem that existing environmental occupancy map construction methods, while aiming to improve query and update efficiency and significantly reduce memory usage, struggle to support real-time, large-scale autonomous navigation tasks in dynamic environments, this application provides a boundary representation-based map construction method and related apparatus. The map employs a global-local map framework. The local map includes the occupancy status of surrounding voxels centered on the robot, while the global map includes the voxel occupancy status in the boundary graph. The voxel occupancy status in the boundary graph is determined by using adjacent voxel pairs to represent the boundary surface layer separating free voxels from adjacent voxels, thus obtaining boundary voxels and constructing the boundary graph. A data structure combining two-dimensional hashing and one-dimensional arrays is used to manage the boundary voxels. Under this data structure, the occupancy status of any voxel is determined through the boundary graph. This application effectively improves query and update efficiency and significantly reduces memory usage.

The present application relates to the technical field of code spraying path planning, and discloses a path planning method, device, equipment and medium for log pile code spraying, which comprises the following steps: collecting position information of the end face of each log in the log pile to be sprayed and corresponding diameter grade information; modeling the log pile to be sprayed into a dynamic layered network model; traversing the log nodes in the dynamic layered network model to generate a code spraying path; performing collision detection on the code spraying path, and optimizing the code spraying path according to the collision detection result; inputting the log code spraying sequence list corresponding to the optimized code spraying path to a code spraying device, and spraying codes according to the log code spraying sequence. The present application converts the log pile into a layered network model by combining the physical constraints of the log pile, traverses the log nodes, plans the code spraying path, embeds the collision detection logic, and dynamically optimizes the code spraying path, thereby replacing manual code spraying, reducing the time complexity of finding the optimal path, and improving the code spraying efficiency.

The invention belongs to the technical field of aircraft target allocation and planning, and particularly relates to a heterogeneous fixed wing cluster target allocation and planning method oriented to flight time collaboration, which comprises the following steps: firstly, introducing a KMEANS method, and clustering targets based on target positions to form target clusters corresponding to unmanned aerial vehicles; then, initial planning of the route in each target cluster is realized based on an improved genetic algorithm; and finally, redistributing and replanning targets in each target cluster on the basis of target position characteristics aiming at the endurance constraint caused by mutual difference of the optimal cruise speeds, so as to realize further reduction of endurance and endurance collaboration. The time-of-flight cooperative target allocation and planning algorithm is low in time complexity, low in computing power requirement and easy to deploy, and can be widely applied to top task planning and target decision allocation tasks of various heterogeneous unmanned aerial vehicle clusters in the future.

This invention discloses a tree-based encoding method and system, belonging to the field of data structure representation technology. The method defines the encoding structure using letters and numbers; it binds encoding relationships by inheriting the parent node's encoding as a prefix for the child node's encoding, and concatenates this prefix with the child node's unique node to form the complete child node encoding; when generating encodings for added nodes, according to the encoding generation logic, it first determines the parent node and then the sibling nodes, and generates encodings in the order of the largest sibling node by sorting the sibling nodes; based on reflection, it calls the encoding function, establishes an Entity base class with necessary attributes and a rule class with generics, and implements the Class class through reflection. <iservice>The initialization of the tree structure. The application reduces the algorithm complexity of the tree structure inquiring all child nodes, and does not need to modify other nodes when adding and deleting, and guarantees the persistence of the database.< / iservice>

The application discloses a kind of program dependence cluster detection methods based on binary set, method is: with binary set form data structure definition dependent set, and input general set operation interface based on bit operation, obtain new dependent set data storage to each instruction node of general set operation interface program, and traverse each instruction node of program;Dependency relationship between instructions is calculated using general set operation interface, and dependency analysis data of each instruction node is obtained;According to the set content in the new set data, the dependency analysis results obtained are reorganized and classified, and the corrected dependency analysis results are obtained to correctly detect the program dependence cluster in program source code.The application defines set data structure with binary set, defines operation interface according to data structure, calculates dependence by calling interface, corrects dependence result, saves the correction result, thereby significantly reduces the space overhead problem during analysis.

The application discloses a minimally invasive fixing device suitable for correction of a concealed penis, which is a non-surgical correction device, comprising a hollow puncture needle, a shrapnel assembly implantable into a human body and a medical suture; a shrapnel push rod capable of pushing the shrapnel assembly out of the puncture needle is arranged in the pipe body of the puncture needle; the shrapnel assembly is arranged in the puncture tip of the puncture needle in a curved shape and is automatically opened after being pushed out by the shrapnel push rod; one end of the medical suture is fixedly connected with the shrapnel assembly, and the other end is led out through the pipe body end of the puncture needle along the length direction of the puncture needle. The shrapnel assembly is bent and accommodated in the puncture tip of the puncture needle, the shrapnel push rod is used for pushing the shrapnel assembly out and automatically opening, and the rapid implantation and anchoring operation under a single puncture path are realized. The conventional general anesthesia correction operation can be replaced, no incision under local anesthesia, rapid correction are realized, and a non-surgical method for the concealed penis is realized, which is local anesthesia, super micro-invasive, rapid correction and safer.

This application provides a method, electronic device, storage medium, and product for updating parameters of a quantum circuit. The method includes: updating a first trainable parameter of a current noisy quantum circuit based on a current policy network to obtain at least one variation path of the first trainable parameter; determining a reward value based on each variation path and a reward function; the accuracy of the reward function and the output of the noisy quantum circuit is related to the degree of influence of noise on the output; updating a second trainable parameter of the policy network based on the reward value; and repeating the above steps until the policy network reaches a preset performance requirement. In this way, the policy network, through reinforcement learning, automatically adjusts the parameterized quantum gates in the noisy quantum circuit to achieve dynamic adaptation of the quantum circuit to noise. By dynamically adjusting the trainable parameters of the quantum circuit through a unified policy network, it is not necessary to model and train separately for each type of noise, thus improving processing efficiency.

The invention relates to a thermal ablation temperature history reconstruction and three-dimensional damage mapping method based on CT thermal imaging end point temperature constraint, and belongs to the technical field of medical image processing and biothermal modeling. The method comprises the following steps: acquiring CT thermal imaging data at a thermal ablation ending moment, and constructing a three-dimensional terminal temperature field of an ablation area; based on the consistency of tissue temperature evolution forms under a fixed ablation clinical protocol, constructing a normalized temperature history template function; reconstructing the discrete time point temperature history of each spatial position in the ablation area by using the temperature history template function; based on the temperature history obtained through reconstruction, a thermal damage evaluation model is introduced for time integration or mapping calculation, and a thermal damage quantized value of each spatial position is obtained; and generating a three-dimensional tissue damage distribution diagram according to the thermal damage quantized value, and visually displaying the three-dimensional tissue damage distribution diagram. According to the method, the quantitative evaluation of the tissue thermal damage in the thermal ablation process can be realized without a real-time temperature monitoring or invasive temperature measuring device.

This invention provides a method, apparatus, device, and storage medium for statistical analysis of embedded data, relating to the fields of data processing and financial technology. The method includes: obtaining a first embedded data array by acquiring first embedded data received at different times; converting the first embedded data array into a second embedded data array by dividing it into preliminary intervals; storing intervals and related information in the second embedded data array; and obtaining statistical values ​​by selecting a subset of array nodes in the second embedded data array based on the statistical time, as required by this invention. By converting the first embedded data array into a second embedded data array, and storing the intervals and information of the first embedded data in the second embedded data array, this invention eliminates the need to traverse the entire array when updating the first embedded data; only the corresponding interval sum needs to be updated to update the second embedded data array, thereby reducing the time complexity during the embedded data supplementation process.

The application belongs to the field of air quality data processing, and particularly relates to an air quality data processing method based on neighborhood rough set attribute reduction. The method comprises the following steps: acquiring air quality sample data with multiple air attributes, and performing permutation and combination on each air attribute; using the dependency of the neighborhood rough set and the calculation of the introduced Shapley value, the importance of each air attribute is calculated, then a weighted neighborhood rough set is constructed according to the importance of the air attribute, the upper and lower approximation relationships of the weighted neighborhood are redefined, then the new dependency is defined, finally the reduced attribute subset greater than the set lower limit of the importance is calculated according to the new dependency; and the air quality data is optimized. The application can solve the problem of consistent air attribute weight in the existing air quality data processing process, and the air attribute obtained through the weighted calculation of the neighborhood rough set can make up for the deficiency of the neighborhood rough set model.

This application provides an image processing method, apparatus, device, medium, and program product, which can be applied in the computer field. The method includes: processing the encoded features of an original image using a super-prior decoding network to obtain the standard deviation and mean of pixel values ​​in multiple dimensions of the original image; determining mask features and inverse mask features based on the standard deviations of the pixel values ​​in multiple dimensions; fusing the mask features and encoded features to obtain target transmission features; performing deterministic sampling on the original image based on the mean of the pixel values ​​in multiple dimensions to obtain sampling features; decoding the pixel positions corresponding to the mean based on the sampling features to obtain the pixels corresponding to the pixel positions; processing the sampling features, inverse mask features, and target transmission features using a decoder network to obtain a compressed image; and storing the compressed image in a target buffer.