83results about How to "Reduce double counting" patented technology

The invention relates to the technical field of performance simulation analysis, in particular to a performance simulation method of a power module. The method comprises the following steps: acquiring initial physical characteristic data and initial electrical characteristic data of a power module; the initial physical characteristic data and the initial electrical characteristic data are preprocessed and simulated, and physical field distribution characteristics in the power module are obtained; performing multi-physical field simulation analysis on the power module based on physical field distribution characteristics in the power module by using an adaptive grid generation algorithm and sparse matrix solution to obtain simulation result data; performing parameter adjustment and optimization on a simulation model of the power module according to the simulation result data to generate optimized simulation model parameter data; and according to the optimization simulation model parameter data, performing simplification processing on the design of the power module to establish a simplified model. Through multi-stage simulation and optimization, the performance and the control precision of the power module can be remarkably improved, and the stability and the efficiency of the system are improved.

The invention discloses a multi-task image processing method, system and device, and relates to the technical field of image processing, and the method comprises the following steps: constructing a multi-task network model; inputting the image data set into a multi-task network model to respectively obtain an infrared enhanced image, an infrared super-resolution image, an infrared and visible light image fusion image and an infrared image colorization image; and calculating a loss function between different images output by the multi-task network model and the infrared image and the visible light image in the input image data set, and carrying out loop training on the multi-task network model. According to the technical scheme, through unified training of the multi-task model, local details are effectively extracted, global consistency is kept, the processing efficiency and the image quality are greatly improved, redundant steps in the processing process are reduced, and the robustness and adaptability of the system are improved.

The present invention discloses a measurement payment method and system based on a BIM model, the method comprising: obtaining a bill of quantities output by a detailed design model; screening out target components that meet preset measurement and pricing conditions from the bill of quantities; extracting price information of the target components from the contract list; matching preset measurement and pricing rules according to the cost attributes of the target components, and generating a payment slip corresponding to the bill of quantities in combination with the price information. The present invention provides a measurement payment method and system based on a BIM model, by establishing a high-precision and high-standard BIM model, applying the BIM model to the bill of quantities calculation, and by extracting the attribute information of the model components, realizing the automatic matching association between the model components and the bill of quantities, which can not only save the calculation time and reduce repeated calculations, but also improve the work efficiency and accuracy of the measurement and pricing stage after quality inspection and evaluation, realize more accurate cost calculation in a three-dimensional model, and reduce the workload of cost personnel.

The invention discloses a large model KV cache rapid storage and import acceleration architecture, and relates to the field of FPGA. The architecture comprises a CPU, an accelerator, an FPGA, a memory, an external memory and a memory. In the query stage, the CPU determines a data query task according to the input lexical elements and controls the FPGA to start the data query task; the FPGA queries and imports model cache data into the memory according to the external memory; in the reasoning stage, the accelerator conducts reasoning and outputs a reasoning result based on the model cache data; and the FPGA reads the updated model cache data from the memory according to a control instruction of the CPU, and exports the updated model cache data to an external memory. According to the scheme, the KV Cache technology is used for reducing the number of times of reasoning repetition of a large model, the hash calculation, table lookup and query access technologies are used for enabling data to be more rapidly imported and exported, and the result output speed is further increased. Not only is the application scene widened, but also data backup and transfer are facilitated.

The invention provides a low-carbon power grid dispatching method based on a hybrid optimization algorithm, and the method specifically comprises the steps: building an optimal dispatching model based on the operation data of a power grid and a carbon emission target; performing preliminary solution based on a dragonfly algorithm to obtain a preliminary global solution of the optimization scheduling model; taking the initial global solution as a starting point of an ant colony algorithm, and performing local optimization through the ant colony algorithm; and monitoring the quality of a current iterative solution of the ant colony algorithm in real time, returning to the dragonfly algorithm when the quality of the current iterative solution meets an interruption condition until the ant colony algorithm can be iterated to meet a corresponding iteration ending condition, outputting a local optimization result, and obtaining a power grid optimization scheduling scheme. According to the method, the multi-objective optimized power grid dispatching can be processed by utilizing a hybrid algorithm consisting of the dragonfly algorithm and the ant colony algorithm, and the convergence speed of optimization iteration of a multi-objective optimized power grid dispatching scheme and the precision of an optimization result are guaranteed.

The invention discloses a power grid business data feature extraction method, which belongs to the technical field of power grids, and comprises the following steps: periodically collecting power grid business data; performing primary processing and feature extraction on the power grid service data to obtain basic feature data; performing feature re-extraction and analysis on the basic feature data to obtain advanced feature data; performing anomaly detection on the basic feature data and the advanced feature data, and judging whether an anomaly detection result is abnormal or not; and when the anomaly detection result is abnormal, triggering an alarm mechanism, and adjusting a feature extraction strategy. According to the power grid business data feature extraction method, multi-level feature extraction and real-time anomaly detection are adopted, and efficient, accurate and real-time power grid business data processing can be achieved.

The present invention relates to the technical field of semantic mapping, and specifically discloses a standardized semantic mapping method and system based on artificial intelligence and big data, comprising: step S1: dividing historical data to be mapped into various sentence segments, and dividing all semantic feature vectors into several classes; recording a class whose total number of semantic feature vectors exceeds the number threshold as a high-frequency class, and selecting a central vector; step S2: comparing the replaceable words of the mapping result and the central mapping result one by one, and obtaining the replacement words of the central mapping result; step S3: storing the central vector of the high-frequency class, all semantic feature vectors, the central mapping result and all replacement words to obtain a mapping library, and obtaining several quasi-mapping results; step S4: selecting the high-frequency class with the highest semantic similarity between the central vector and the semantic vector, obtaining the quasi-mapping result corresponding to the semantic feature vector with the highest similarity to the semantic vector in the high-frequency class, and using it as the mapping result of the sentence segment to be mapped.

The invention discloses a multi-modal disambiguation and knowledge graph embedding-based power grid Internet of Things equipment plug and play method and system, and the method comprises the steps: analyzing an ambiguous user instruction through a multi-modal module, and fusing context and sensor data to generate a clear configuration intention; a knowledge graph containing equipment configuration and topological resources is constructed, after the knowledge graph is converted into a vector through embedding, an optimal access scheme and a physical model are matched through machine learning, meanwhile, power grid operation rule verification is conducted on a matching result, and it is ensured that the scheme can be directly executed. According to the method, the problems of inconsistent object model matching, manual configuration migration, weak ambiguous instruction analysis and the like are solved, automatic access of novel switches, photovoltaic equipment and other equipment is realized, and the automation level of the power grid internet of things is improved.

The invention discloses a task scheduling method, electronic equipment and a computer program product, and relates to the technical field of data processing.The method comprises the steps that reusable procedures are determined from data processing tasks to be scheduled, the data processing tasks are composed of at least one atomization procedure, and the reusable procedures are reusable atomization procedures; and scheduling the data processing task through a dynamic planning algorithm, and multiplexing the reusable process in the scheduling process. The resource utilization rate and the task scheduling efficiency in the complex data processing task scheduling process can be improved.

The invention relates to the technical field of databases, in particular to a data processing method, device and equipment. A data processing method comprises the steps that query statements in workflows to be processed are analyzed, keyword information of the workflows is obtained, and the keyword information comprises target table fields, query result fields, source table fields, query condition fields, grouping fields and sorting fields of the workflows; performing optimization processing on each workflow according to the keyword information of each workflow; and updating the data warehouse according to the optimization processing result of the workflow. Through the embodiment of the invention, the repeated calculation of the workflow is reduced, the calculation resources and the calculation cost are saved, the reasonability of the data processing task is improved, the maintainability is improved, the task robustness is improved, and the operation and maintenance cost is saved.

The invention discloses a data processing method and device, equipment and a storage medium, and relates to the technical field of data processing. The method comprises the steps of obtaining a first mode sequence related to a network structure optimization task, wherein the number of network nodes corresponding to a target mode representing the same constraint condition in at least two service rules is lower than a preset number; according to the evaluation excitation degree of the first mode sequence, a second mode sequence is extracted from the first mode sequence set, and the evaluation excitation degree is determined by the affinity of the matching degree of the characterization mode sequence and the network structure optimization task and the antibody concentration of the similarity degree of the characterization mode sequence and other mode sequences in the mode sequence set where the characterization mode sequence is located; screening a third mode sequence of which the evaluation excitation degree is lower than a preset threshold value from a clone deformation mode sequence set of the ith second mode sequence in the second mode sequences; the Rete network constructed through the third mode sequence meets the indication of the network structure optimization task. Therefore, the Rete network structure is simplified, and the memory occupancy rate is reduced.

The invention is suitable for the technical field of GPU (Graphics Processing Unit), and provides an SPIR-V-based inter-thread shared memory processing method, which comprises the following steps of: S1, reading a high-level language code source code to a character buffer area, and then converting the source code in the character buffer area into an intermediate code in an SPIR-V form; s2, traversing an instruction sequence of an intermediate code, and for variable declarations in the instruction sequence, accumulating the sum of occupied spaces of all shared memory variables; s3, calculating a one-dimensional index and a shared memory address of the thread group; and S4, traversing the instruction sequence of the intermediate code, and for instructions in the function declaration and the function definition, if processing of a shared memory is involved, calculating a real address and generating different GPU instructions according to instruction types. The invention provides a universal method for translating related codes, used for processing a shared memory of a working group, in SPIR-V codes into basic instructions of all GPU architectures, and the instructions used in the method are basic instructions of all GPU instruction sets, so that the method has hardware platform universality.

The invention discloses a method and system for constructing a comprehensive evaluation index of a soil water source conservation function, and relates to the technical field of soil water source conservation, and the method comprises the following steps: measuring a soil water retention performance index and a soil infiltration performance index; based on the measured indexes, determining the weight of each index by adopting a factor analysis method, and carrying out membership standardization processing on each index; and weighting the weight of each index and the standardized membership degree to obtain the comprehensive evaluation index of the soil water conservation function. According to the method, the soil water source conservation function can be quantitatively evaluated on the sample plot scale only through one-time or few-time soil index determination, dependence on a long-term hydrological sequence is avoided, and the data acquisition cost is remarkably reduced.

The embodiment of the invention discloses a gas diffusion translation calculation method and device, computer equipment and a storage medium. The method comprises the following steps: setting gas diffusion related parameters of each standard puff; initializing an observation space, an observation duration and an observation time point, and calculating a downwind direction and a crosswind direction; calculating a concentration matrix sequence of the standard puff of each ship at each time point within the observation duration to obtain a concentration matrix corresponding to each ship; and calculating a puff group concentration matrix of a plurality of ships at a specified observation point in the observation space according to the concentration matrix by adopting a standard puff multi-translation multiplexing mode. By implementing the method provided by the embodiment of the invention, the algorithm complexity during multi-ship continuous emission calculation can be greatly simplified, the repeated calculation amount is reduced, the memory space occupation is reduced, and the calculation time consumption is reduced.

The invention relates to a content generation method combining a knowledge graph and a large language model. The method comprises the following steps: based on multi-type target content data, extracting an entity, relationship and attribute triple of a multi-source corpus, dividing a theme cluster, constructing a knowledge graph, and carrying out standardized packaging on knowledge units to obtain a standardized knowledge set; analyzing a user input demand, obtaining a multi-dimensional demand feature, performing similarity matching with a theme cluster to obtain a target cluster, screening candidate knowledge through multi-hop retrieval, and obtaining an associated knowledge subset through three-dimensional evaluation index empowerment score sorting; converting the associated knowledge subset into an analyzable structured text, and adjusting and optimizing model generation parameters to obtain initial generation content; and comparing the initial content with a knowledge graph to obtain a result, optimizing text features and performing adaptive adjustment in combination with multi-type target content data and audience features, and obtaining standard content of target quality. Content generation in combination with the knowledge graph and the large language model is realized.

The embodiment of the invention discloses a communication signal multi-parameter intelligent parallel extraction method. The method comprises the following steps: firstly, acquiring a to-be-measured signal and extracting an I / Q component; processing the I / Q components through a multi-scale feature extraction structure to obtain high-dimensional features; extracting shared features from the high-dimensional features through a plurality of residual block groups, inputting a plurality of parallel task branches, and enhancing the shared features in the branches by using a channel attention mechanism to obtain output features; the output features of all task branches are transformed through a learnable linear layer to form a key matrix and a value matrix, the current task features are transformed through a learnable linear layer to form a query matrix, and the input features are segmented into a plurality of heads by means of a multi-head attention mechanism; each head obtains output of a single head based on zoom dot product self-attention operation, and results of all the heads are integrated to obtain enhanced features of the current task; and determining a communication signal parameter prediction label according to the enhanced features, and realizing multi-parameter efficient parallel extraction.

The application provides a vulnerability repair method and system, and relates to the technical field of network security vulnerability verification and enterprise vulnerability management. The application obtains a task containing a vulnerability identifier, asset information and to-be-verified repair content, processes the task in parallel according to a pipeline split according to business execution steps, and configures a resource pool matching external dependencies for each stage; a stage fingerprint is generated when the task is transferred, a result is reused when a subtask is executed, otherwise the execution is scheduled according to vulnerability harm, asset importance and threat intelligence, and the vulnerability repair is completed according to the verification result. The application solves the problems of low throughput of the existing general pipeline, slow response of high-priority tasks, and easy starvation of low-priority tasks, and achieves the effect of collaborative optimization of throughput improvement, emergency task response acceleration and low-priority task starvation prevention without changing the original business and hardware.

The present invention provides a method for layered processing of homomorphically encrypted data and a readable storage medium, comprising: dividing the homomorphically encrypted data into a plurality of sub-data according to the attributes; storing each sub-data in a separate storage space to form a plurality of data layers; each data layer corresponds to a cache layer, and after processing the sub-data of any data layer each time, the processing result is stored in the cache layer corresponding to the data layer to form cache data; according to the business requirements for the whole or partial processing of the homomorphically encrypted data, a plurality of the data layers are selected for processing to obtain the business processing sub-result of the data layer, and when the cached data is required by the business requirements, the cached data is directly called as the business processing sub-result of the data layer corresponding to the cached data and the data processing of the data layer corresponding to the cached data is skipped. The present invention can realize layered processing and efficient calling of data.

The invention discloses a text classification method and a storage medium, and the method comprises the steps: obtaining a data set constructed by a to-be-classified text, constructing a graph structure based on the data set, and obtaining an adjacent matrix generated by the graph structure; pre-training nodes in the graph structure to obtain a node vector adapted to the graph structure; inputting the feature matrix of the node vector and the adjacent matrix into a graph convolutional network, and classifying the text; wherein the feature matrix of the node vector and the adjacent matrix are input into the graph convolutional network, and the text is classified, which comprises the following steps: taking the feature matrix as input, and in each layer of the graph convolutional network, using the input feature matrix of the previous layer to obtain the hidden layer representation of the next layer; and mapping the output of the last layer of the graph convolutional network to a result space through a feedforward network to obtain the final output of each node. According to the method, the prediction capability of the unseen nodes is improved, and the reasoning efficiency on a large-scale graph data set is improved.

The invention relates to a positioning method and device, electronic equipment and a storage medium. The method comprises the steps that a first positioning signal is received, the first positioning signal is sent when a sending device detects that the first positioning signal is different from a second positioning signal, and the second positioning signal is a positioning signal sent by the sending device last time; and performing positioning processing based on the first positioning signal to obtain position information. According to the positioning mode, the latest positioning signals can be received in time, the obtained position information has high real-time performance, real-time positioning is achieved, the position information does not need to be repeatedly obtained when the positioning signals are not changed, and therefore unnecessary repeated calculation is reduced.

The present invention discloses a large language reasoning system, including: at least one pre-filling calculation unit, at least one generating calculation unit and a cache unit; at least one pre-filling calculation unit is used to pre-fill according to the current reasoning request and historical reasoning resources, obtain the current reasoning data, and send the current reasoning data to the generating calculation unit and the cache unit; at least one generating calculation unit is used to receive the current reasoning data, and use a preset large language model to generate the current result word according to the current reasoning data, generate the current reasoning result according to the current result word and send it to the cache unit, and output the current result word to the user end; the cache unit is used to save the current reasoning data, the current reasoning request and the current reasoning result according to the current reasoning request. The technical solution provided by the present invention can reduce the amount of calculation in the reasoning process and improve the fluency and continuity of the natural language of the reasoning result.

The invention discloses a voice text bidirectional conversion system and method based on asynchronous processing and a message queue. The system comprises a server, an OBS object storage service, a business system, a conversion gateway and a plurality of universal gateways. The conversion gateway is connected with the server, a voice-to-text module and a text-to-voice module are configured in the conversion gateway, and the voice-to-text module is constructed based on a FunASR streaming recognition framework, performs communication through a WebSocket protocol, and performs a voice recognition process in combination with an asynchronous task processing mechanism; the text-to-voice module is constructed based on ChatTTS service, and realizes an asynchronous task distribution and result return mechanism in combination with a RabbitMQ message queue. According to the scheme, efficient conversion between the voice and the text is realized through unified interface design and modular processing flow, and the response speed, the accuracy and the expandability of the voice processing system are improved.

The present application provides a watershed extraction method based on outlet relocation, including: determining a buffer zone based on a given outlet, iteratively calculating the local cumulative flow in the buffer zone according to the flow direction data and outputting the global cumulative flow; determining the main stream network based on the global cumulative flow and flow direction data, relocating the given outlet coordinates to the main stream network; relocating the outlet coordinates to the tributaries of the main stream; determining the reference outlet according to the main stream relocation results and the tributary relocation results, and extracting the watershed according to the reference outlet. The outlet relocation algorithm in the present application is insensitive to preset parameters, suitable for mass extraction of watersheds of different areas, and only uses the flow direction and given outlet coordinates as input. The algorithm provides a tool for automatically extracting a large number of watersheds for hydrological simulation in the big data era.

The invention discloses a batch data processing method and device suitable for a database, equipment and a medium, relates to the technical field of data processing, and is applied to a processor. Monitoring whether a target instruction meeting a preset stride condition exists in a data processing instruction stream corresponding to the preset database in the target assembly line or not, if yes, entering an instruction pre-execution mode, vectorizing the target instruction and a subsequent instruction having a dependency relationship with a register of the target instruction, and outputting the vectorized instruction; performing instruction pre-execution based on the vectorized instruction and a first execution unit, and storing an instruction pre-execution result to a first physical register file and a result cache region; and when the data processing instruction stream in the target assembly line reaches the input end of the second execution unit, matching the instruction pre-execution result in the result cache region, and completing batch data processing operation based on an information matching result. According to the invention, the efficiency and accuracy of batch data processing are improved.

The present invention relates to a die-casting process simulation method based on cloud computing, which belongs to the technical field of die-casting process, and comprises the following steps: step one: according to the die-casting process, the die-casting process parameters are input into the simulation software to construct an equation for simulating molten metal filling; step two: the simulation software simulates the process of metal filling according to the equation, generates a result of filling rate changing with time, and uploads it to a cloud platform; step three: the cloud platform calculates the difference between the generated result and the actual filling process, judges whether the difference is greater than a threshold value, and when the judgment result is yes, calculates a critical solid phase fraction correction value, and executes step four, and when the judgment result is no, outputs the critical solid phase fraction; step four: the cloud platform sends the critical solid phase fraction correction value to the simulation software, and the simulation software corrects it according to the critical solid phase fraction, and then generates a new result of filling rate changing with time according to the corrected equation for simulating molten metal filling, and uploads it to the cloud platform, and returns to step three.

The invention provides a font recognition method, device and equipment based on self-adaptive preprocessing and a medium, and the method comprises the steps: carrying out the self-adaptive preprocessing of a collected font image, the preprocessing comprises dynamic noise suppression, illumination unevenness correction and sub-pixel-level geometric correction, and obtaining a corrected character image; performing multi-dimensional feature extraction on the corrected character image, and constructing a fusion feature vector including local detail features, global style features and geometric correction features; and performing dynamic weight matching based on the feature contribution degree evaluation model, calculating the similarity between the to-be-matched character and the candidate fonts, and outputting the recognized font type according to the similarity result, thereby ensuring the font recognition accuracy.

The present invention provides a grayscale release migration method and system for an information system, which relates to the field of information system migration. The method controls the grayscale ratio of each module by initializing a grayscale ratio vector, dynamically diverts user requests; monitors the abnormality and slow response of the system with migration based on time slices, and generates a system statistical vector; performs weighted statistics and standardization on the statistical vector using a sliding window to form a system state vector; combines health status information to generate a final system state vector reflecting the real-time operating status of the system; calculates the stability index of each module through a stability discrimination model, dynamically adjusts the grayscale ratio, prioritizes the transition of stable modules, and rolls back the grayscale ratio of unstable modules. The system supports automatic adjustment of grayscale ratios, real-time risk response, and multi-architecture compatibility, significantly reduces the risk of the migration process of the information system, and improves the efficiency and stability of grayscale release.

The invention provides a view multiplexing component development method for realizing local rendering on a guLiteOS platform, and aims to optimize resource utilization and improve system response speed through dynamic loading, event-driven local updating and a view multiplexing technology. The method comprises the following steps of: firstly, dynamically loading a core operation component according to user interaction and a system state, marking a non-core component as a'to-be-loaded 'state, and delaying loading until a specific demand is triggered; and secondly, through an event-driven mechanism, user operation is captured, an influence area is calculated, and only an influenced interface part is locally rendered, so that resource waste caused by global rendering is avoided. And finally, through a view multiplexing mechanism, storing a component state after first rendering, and dynamically adjusting a caching and multiplexing strategy according to user operation frequency and importance. According to the method, repeated rendering can be reduced, memory use is optimized, and the overall performance of the system and the user experience are improved.

According to the static stain analysis device and analysis method based on semantic harmless recognition, a traditional data flow analysis framework is abandoned, a corresponding pollution propagation rule is formulated, and a pollution path diagram related to a pollution variable relation is abstracted; and a user-defined Sanitizer set is recognized based on semantics, so that the pollution path diagram is further verified and pruned through the Sanitizer set. According to the analysis method, a corresponding pollution path diagram is constructed through a defined pollution propagation rule analysis program, and the device is only focused in a data stream of a stain source and a leakage point, so that the time cost for analyzing redundant nodes and edges is reduced. According to the method, the detection capability and analysis efficiency of static taint analysis on vulnerabilities in the Web application program can be improved, and more accurate input data is provided for subsequent security analysis.

The invention belongs to the technical field of quantum computing, and particularly relates to a quantum circuit simulation method based on circuit cutting. By designing a cutting strategy, a complex quantum circuit is divided into a plurality of small sub-circuits, and each sub-circuit can be independently simulated; the method specifically comprises the steps of circuit preprocessing and dependency relationship analysis; sub-circuits are divided through circuit cutting, specifically, a heuristic cutting algorithm is adopted, and an optimal cutting point is determined according to the dependency relationship between the circuits; selecting a scheme capable of minimizing the overall simulation burden by evaluating the overhead of different cutting schemes to optimize the global calculation cost and resource allocation so as to ensure that the cut sub-circuits can be independently and efficiently simulated; and sub-circuit state vector multiplexing: storing the intermediate state vector of each sub-circuit, and searching similar state vectors for multiplexing when a new sub-circuit is calculated, so as to reduce the calculation amount and resource consumption. According to the method, the complexity of large-scale quantum circuit simulation can be reduced, and the calculation efficiency is improved.