178 results about "Preprocessor" patented technology

The invention discloses a power plant operation and maintenance knowledge intelligent query method based on a large language model and an RAG technology. The method comprises the following steps: constructing a power plant operation and maintenance knowledge vector library covering structured, semi-structured and unstructured data; receiving a natural language question of a user, inputting an improved instruction to align a preprocessor, and generating a question semantic vector and an intention tag; relevant knowledge fragments are retrieved and sorted through a semantic matching retriever in combination with the intention labels; constructing a large language model cue word structure based on the retrieval result and the original question, generating candidate answers and recording a reference path; and finally, performing term specification and consistency verification according to the expert rule base, and outputting a structured and traceable final answer. According to the invention, the improved RAG technology is fused to realize intelligent query of the operation and maintenance knowledge of the power plant.

A deep learning system for time series prediction comprising a preprocessor that receives time series input sequences, truncates them by removing terminal values, and appends padding values to maintain the original sequence length. An encoder compresses these padded sequences into latent space representations, while a decoder reconstructs predicted sequences matching the original length, specifically trained to reconstruct values matching the removed terminal values in positions corresponding to the padding values. A training system optimizes the encoder and decoder by minimizing differences between original sequences and predicted sequences. The system can process multiple time horizons simultaneously while maintaining statistical properties and providing uncertainty quantification through confidence intervals. This approach enables accurate short-term forecasting while preserving both temporal patterns and statistical relationships in the predicted sequences.

An adaptive data system (ADS) for cognitive data processing is disclosed. The ADS includes an adaptive semantic preprocessor, a trigger detector, a temporal batching engine, a symbolic encoder, and a dynamic cognitive transformer engine. The adaptive semantic preprocessor is configured to receive input data from one or more databases and identify cognitive data attributes comprising one or more contextual, semantic, and temporal attributes from the received input data. The trigger detector is configured to identify semantic divergence of the identified cognitive data attributes and provide a standardized data. The temporal batching engine is configured to provide a high-dimensional cognitive data from the standardized data. The symbolic encoder compresses the high-dimensional cognitive data. The dynamic cognitive transformer engine is configured to determine decision making rules, analyze the compressed high-dimensional cognitive data based on the decision making rules and provide recommendations based on an outcome of the analysis to a user.

A networked data analytics management (DAM) system includes a data representation manager (DRM), a data collection and preprocessor (DCP), a correlation manager (CM) and a data source discoverer (DSD). The DRM obtains representations of data sources, used to represent data source characteristic. The DCP interacts with the data sources to obtain data via data collection actions. The CM determines correlations between different obtainable data, based on the representations. The DSD interacts with a data consumer and the CM to determine the set of data collection actions to perform, and to configure the DCP.

Systems and methods for identifying events in large sums of data using large language models. A system includes a data shipper configured to ingest raw data and a data preprocessor configured to receive the raw data from the data shipper and processes the raw data to generate processed data. The system includes a database that stores the processed data and a machine learning engine in communication with the database. The machine learning engine executes a large language model algorithm on the processed data to identify one or more of an anomaly in the processed data or two or more correlated events in the processed data.

The invention provides a sparse approximate inverse Shenwei parallel preprocessing method and system for electromagnetic scattering, and relates to the technical field of processor parallel computing, and the method comprises the steps: carrying out the performance hotspot analysis of a sparse approximate inverse preprocessor in an electromagnetic scattering simulation process, and determining a hotspot function of the sparse approximate inverse preprocessor; extracting a sparse matrix vector multiplication operation of a hotspot function, performing master-slave core parallel calculation on the sparse matrix vector multiplication operation, migrating the sparse matrix vector multiplication operation to slave core calculation, and accelerating hotspot calculation by using a heterogeneous parallel sparse approximate inverse algorithm; parallel computing of master-slave core intensive numerical computing tasks is achieved; in the parallel computing process, the slave cores communicate with the master core in a direct memory access mode, and during the period, the master core is in a waiting state to ensure data consistency and communication synchronization until all the slave cores compute distributed tasks.

A method for classifying neurological disease status is described. The method includes acquiring, by a data preprocessor logic, patient image data. The method further includes generating, by a trained artificial neural network (ANN), a classification output based, at least in part, on the patient image data. The classification output corresponds to a neurological disease status of the patient. The trained ANN is trained based, at least in part, on longitudinal source data.

Systems and methods for detecting tampering in software are disclosed. The system includes a preprocessor that converts source code into a minimal intermediate representation and extracts semantic and syntactic features using word embedding algorithms. The preprocessed data is then fed into two machine learning models: a classical LSTM model and a quantum LSTM model. The classical LSTM model detects basic tampering patterns, while the QLSTM model leverages quantum principles to enhance analysis and prediction of more complex tampering attempts. The system also includes a quantum cache for efficient data retrieval and manipulation, enabling real-time or near-real-time analysis. The combination of these features provides improved accuracy and effectiveness in detecting tampering, enabling timely intervention and mitigation of security threats. Remediation may be performed automatically or manually and can be based on historically determined or dynamically generated solutions.

The invention belongs to an electromagnetic simulation technology, and particularly provides an electromagnetic scattering simulation method based on a skeleton decomposition precondition, which is used for completing electromagnetic scattering simulation calculation of an electrically large complex target, improving the ill-conditioned problem of a complex target generation impedance matrix while ensuring the simulation calculation precision, and improving the convergence and reliability of an iteration method. According to the method, firstly, a bone element primary function is selected by using an interpolation decomposition technology and a proxy surface method, and an approximate inverse matrix of an impedance matrix is obtained by using a skeleton decomposition method; then, the approximate inverse matrix is adopted as a pre-processor, a current coefficient is obtained through a Krylov subspace iteration method, and finally electromagnetic scattering of the metal target is obtained through calculation according to the current coefficient. The invention creatively provides a brand-new pre-processor, and compared with the traditional pre-processor, the pre-processor can further improve the matrix behavior, greatly reduce the iteration time and improve the convergence.

The invention provides a storage and calculation integrated neural mimicry hardware-oriented spiking neural network compiling system and method, and the system comprises a preprocessor part which employs a heuristic algorithm to convert a conventional artificial neural network in an input model into a spiking neural network, and carries out the compiling of an ANN-SNN hybrid network into a computational graph and a neural topology network; and the compiler part is used for compiling a computational graph of the ANN-SNN hybrid network to generate a binary file which can be executed by the storage and calculation integrated neural mimicry hardware, and the compiling stage comprises three universal compiling stages, namely a neuron division stage, a core mapping stage and an instruction generation stage. According to the invention, the problem of storage and calculation integrated neural mimicry hardware acceleration of the ANN-SNN hybrid network is solved, and it is ensured that the ANN-SNN hybrid network can be compiled and submitted to the storage and calculation integrated neural mimicry hardware for acceleration.

The invention relates to the technical field of image processing, and discloses an image rain removal method and device, equipment and a storage medium, which are used for improving the image rain removal precision and efficiency. The image rain removal method comprises the following steps: inputting an original rain image into a data preprocessor of a pre-trained target image rain removal model for preprocessing to obtain a candidate rain image; inputting the candidate rain map into a feature extractor, and performing feature extraction through feature extraction layers of a plurality of stages which are connected in sequence to obtain a multi-scale feature map; inputting the multi-scale feature map into a decoder, segmenting the multi-scale feature map through a first feature pyramid network propagating from bottom to top in a cascade feature pyramid fusion module, and processing the initial segmented feature map of each layer through a pyramid pooling module and a multi-scale large-selection kernel attention module; performing segmentation through a second feature pyramid network propagating downwards from the top layer to obtain a target segmentation feature map of each layer; and performing prediction based on the target fusion feature map to obtain a target rain-free map.

Provided are a memory mapping method and system for a rotary position embedding operation. The memory mapping system according to one embodiment may include: a preprocessing unit for adjusting the positions of the values in a weight matrix for a pretrained artificial intelligence model such that the values of the elements required in a rotary position embedding operation, which is the next operation after a matrix multiplication operation, are adjacent to one another; a weight memory for storing the weight matrix with the positions of the values adjusted; and a hardware operator for generating a query vector or a key vector by processing the matrix multiplication operation of an input vector and the weight matrix and for processing the rotary position embedding operation of the query vector or the key vector by using the values of the elements adjacent to one another.

In one embodiment, a method includes receiving data of a set of configurations of preprocessor engines, receiving measurements of performance of a device executing benchmark applications while changing a configuration of preprocessor engines selected from the set of configurations of preprocessor engines, defining an order of at least some of the configurations based on the measurements, and providing a pruned set of configurations based on the defined order of the at least some configurations.

In one embodiment, a method includes receiving data of a set of configurations of preprocessor engines, receiving measurements of performance of a device executing benchmark applications while changing a configuration of preprocessor engines selected from the set of configurations of preprocessor engines, defining an order of at least some of the configurations based on the measurements, and providing a pruned set of configurations based on the defined order of the at least some configurations.

A data storage device includes a memory array for storing data; a host interface for providing an interface with a host computer running an application; a central control unit configured to receive a command in a submission queue from the application and initiate a search process in response to a search query command; a preprocessor configured to reformat data contained in the search query command and generate a reformatted data; and one or more data processing units configured to extract one or more features from the reformatted data and perform a data operation on the data stored in the memory array in response to the search query command and return matching data from the data stored in the memory array to the application via the host interface.

The invention relates to an automobile connector production surface quality detection method, and relates to the technical field of automobile connector surface detection, and the method comprises the steps: obtaining a connector surface image through an image collection device; the image preprocessor performs affine transformation on the image to generate a transformed image, and extracts a detail layer image; the filtering unit acts on the detail layer image according to the filtering kernel to generate a filtering image; positioning and determining a suspected abnormal region; combining the two suspected areas to determine a final abnormal range; and the calculation unit constructs a Gaussian distribution model according to the non-abnormal pixel points, and judges whether the connector is qualified or not through comparison with an abnormal region, so that the detection rate is remarkably improved, and meanwhile, relatively high recognition accuracy is kept. And complicated mathematical operation is applied to the image sequence, so that the defect that general static analysis is easily interfered by external noise is overcome, and the overall robustness of the algorithm is enhanced.

A preprocessor for use with a machine learning system for control of computerized avatars provides for an embedding of avatar control information in a speech response file machine learning system for improved perception of emotional intelligence.

A method and system for analyzing a vehicle winch design includes a preprocessor determining contact locations for winch components from a mesh model. The preprocessor determines load vectors at a winch assembly and counter-vectors at a fairlead based on a relative position of a winch wire and a hook retention location based on a winch load rating. A non-linear analysis system determines plastic strains, deflections, and clearances under the winch load rating based on the contact locations, load vector and counter vector, and communicates the plastic strains, deflections, and clearances to a post processing system. The post processing system compares the plastic strains to strain limit at a post processor, compares the deflection to a deflection limit, compares clearances to a clearance limit and generating a display based on comparing the plastic strains, compares the deflection to the deflection limit and compares clearances to the clearance limit.

The invention discloses an API interface development method and device, equipment and a storage medium, and relates to the technical field of software development, and the method comprises the following steps: determining API interface definition and interface implementation, storing API interface definition information and interface implementation in an interface definition table and an interface implementation table, determining a binding relationship between the API interface definition and the interface implementation, and storing the binding relationship in a binding relationship table; determining a target HTTP request, querying an API interface definition based on a request path, determining target interface implementation, and loading the target interface implementation to a memory; calling a preprocessor to process an original request parameter of the target HTTP request to obtain a processed request parameter, and executing processing logic on the processed request parameter based on the implementation mode to obtain an execution result; and processing the execution result by using a rear processor to obtain a target processing result, and obtaining an HTTP response object according to the target processing result and sending the HTTP response object to the calling party. The flexibility and expansibility of the interface can be improved.

The invention relates to the technical field of artificial intelligence, and discloses a separated multi-mode large language model service system and a first lexical element generation method. The system comprises a preprocessor, an encoder instance, an instance interaction layer and a pre-filling instance, the preprocessor is configured to analyze the question and answer request to obtain original data of multiple modals; the encoder instance is used for carrying out parallel encoding on the original data of each mode to generate a subsequence of the corresponding mode; the instance interaction layer is configured to send a sub-sequence of each mode generated by the encoder instance to the pre-filling instance; the pre-filling instance comprises a large language model trunk and is configured to perform asynchronous pre-filling on each sub-sequence of each mode by taking the sub-sequence as granularity so as to generate a first lexical element of an answer text corresponding to the question and answer request. By adopting the system, the response speed of the online question and answer service system can be improved, and the user experience is improved.

The invention discloses an epileptic seizure detection system based on a multi-dimensional hypergraph fusion network, which comprises a multi-modal data collector, a signal preprocessor, a feature extractor, a multi-dimensional hypergraph builder, a hypergraph feature extractor, a feature fusion device, a classifier and a time sequence prediction corrector, the multi-dimensional hypergraph constructor constructs three hypergraph structures, namely an intra-modal hypergraph, an inter-modal hypergraph and a time sequence hypergraph; the feature fusion device integrates the features extracted by the three hypergraphs to generate a comprehensive feature vector; the classifier classifies the fusion features through a two-layer feedforward neural network, and outputs probability prediction of epileptic seizure; the time sequence prediction corrector applies a majority voting smoothing filter to carry out post-processing on a classification result, isolated misclassification is eliminated, and the time sequence consistency of prediction is enhanced. The invention provides an innovative multi-dimensional hypergraph fusion network, and the robustness and accuracy of epileptic seizure detection are remarkably improved by simultaneously modeling high-order relationships in modals, between modals and time sequence dimensions.

The invention discloses a multi-source heterogeneous big data driven intelligent contract abnormal behavior real-time detection and early warning method, and relates to the technical field of block chains. According to the invention, a multilayer data preprocessor is adopted to carry out standardized conversion on intelligent contract codes, transaction records and network traffic, the problem that the traditional method is limited to single data source analysis so that the detection accuracy is low is solved, and a graph neural network, a time sequence analysis algorithm and a deep auto-encoder are utilized to carry out deep analysis on data. Multi-dimensional features of abnormal behaviors are comprehensively captured, the detection accuracy is remarkably improved, false alarm and missing alarm are effectively reduced, real-time data streams are processed by means of a streaming computing framework, millisecond-level real-time detection, quick response and early warning of the abnormal behaviors are achieved through parallel feature extraction and distributed computing resource scheduling, and safe operation of smart contracts is ensured.

The invention relates to the technical field of ion beam physical calculation, in particular to an ion beam optical property evaluation method based on a CPU-GPU hybrid parallel stable double-gradient conjugate algorithm. According to the technical scheme, the method comprises the following steps that a CPU-GPU hybrid parallel architecture is initialized, and CPU end and GPU end memory allocation, CUDA related object creation and GPU equipment initialization and parameter setting are completed; constructing a linear equation set, a sparse coefficient matrix A and a right-end vector b, and setting an initial solution vector, convergence precision epsilon and related parameters of the number of iterations; and analyzing the sparseness of the coefficient matrix A. Through combination of three core mechanisms of CPU-GPU cooperative calculation, dynamic load balancing and intelligent preprocessor selection, efficient, stable and universal evaluation of the ion beam optical characteristics is successfully realized, and an excellent solution is provided for solving the solving problem of a large-scale sparse linear equation set in the field of high-performance calculation.

The present invention provides an industrial equipment fault diagnosis method, system, and storage medium based on multi-view expansion statistical features, which relate to the technical field of industrial equipment fault diagnosis. The method aims to solve the problems that existing methods are easily contaminated by noise, and that the extracted features contain redundant and irrelevant features, resulting in long calculation time, low calculation efficiency, and high hardware computing power requirements. The method comprises the following steps: step 1, obtaining a time series data set of the operating fault status of the industrial equipment; step 2, preprocessing the time series data using an unsupervised differential expansion mapping preprocessor; step 3, constructing multiple filtering feature selectors at different storage ratios to obtain multi-view feature subset vectors and a combined feature selector; step 4, constructing multiple classifiers, training the multiple classifiers based on the multi-view feature subsets, and constructing a combined classifier; and step 5, using the combined feature selector and the combined classifier to perform fault diagnosis on the equipment operating status.

An activity classifier system and method that classifies human activities using 2D skeleton data. The system includes a skeleton preprocessor that transforms the 2D skeleton data into transformed skeleton data, the transformed skeleton data comprising scaled, relative joint positions and relative joint velocities. The system also includes a gesture classifier comprising a first recurrent neural network that receives the transformed skeleton data, and is trained to identify the most probable of a plurality of gestures. The system also has an action classifier comprising a second recurrent neural network that receives information from the first recurrent neural networks and is trained to identify the most probable of a plurality of actions.

A method for processing an input frame for an on-device AI model is provided. The method may include obtaining an input frame. The method may include building at least one kernel independent of the scale of the input frame by passing input variables to the at least one kernel using preprocessor directives independent of the scale of the input frame. The method may include inputting the input frame to the on-device AI model including the at least one kernel independent of the scale of the input frame. The method may include processing the input frame in the on-device AI model.

A multidimensional knowledge transfer model for predicting and calculating ad click-through rates (CTRs). The multidimensional knowledge transfer model includes a preprocessor for constructing an ad group node graph based on similarities between ad group nodes, constructing an ad campaign node graph by merging ad group node graphs, and constructing an ad account node graph by merging ad campaign node graphs. The multidimensional knowledge transfer model also includes a multi-knowledge click-through rate (CTR) prediction model for each layer: the ad account, ad campaign, and ad group layers. The multi-knowledge CTR prediction model predicts the CTR of each node based on the ad account node graph, ad campaign node graph, or ad group node graph, audience characteristics, characteristics of nodes for which CTRs have been predicted, and the hidden vector of its parent node extracted from the upper-layer multi-knowledge CTR prediction model.

The invention relates to a business data processing platform, a data processing method, an electronic device and a storage medium, in the business data processing platform, a first preprocessor is used for responding to a target bookkeeping request and matching a corresponding first target rule from a rule base; the accounting system is used for updating account information of a first account corresponding to the target accounting request according to the first target rule, and generating a first accounting detail corresponding to the target accounting request to obtain a first accounting processing result; pushing the first accounting processing result and the message to an accounting system; the second preprocessor is used for matching a corresponding second target rule from the rule base; and the accounting system is used for updating account information of a second account corresponding to the target accounting request and generating a second accounting detail corresponding to the target accounting request. The processing logic of the accounting system does not need to be changed frequently according to business changes, and the flexibility and expandability of data processing of the accounting system are improved.

The invention discloses a quantum linear solution method, device and equipment based on a polynomial preprocessor. The method comprises: obtaining information of a first matrix A and a first vector b in a linear system to be processed, calculating a polynomial p(A) of the first matrix A used for preprocessing the linear system, preprocessing the linear system according to the polynomial p(A), obtaining a second matrix A′ and a second vector b′, constructing a quantum circuit corresponding to an HHL algorithm, performing evolution and measurement operations of a quantum state according to the second matrix A′ and the second vector b′, and obtaining the final quantum state of the evolved quantum circuit, so as to solve the deficiencies in the prior art. The method can reduce the time complexity and the amount of calculation for solving linear problems, and at the same time reduce the occupation of hardware resources.