56 results about "Code algorithm" patented technology

The invention discloses a sprint learning path optimization method based on an improved whale optimization algorithm. The problem that existing sprint path planning is unreliable is solved. The method comprises the following steps: constructing a total learning cost objective function; using a random key coding algorithm based on chaotic mapping to represent the learning paths in a sequence mode, and using serial scheduling to obtain a plurality of candidate sprint learning paths; introducing an adaptive weight mechanism, an examination driving offset item and a Gaussian disturbance item, and adopting an improved white whale optimization algorithm to carry out iterative optimization on the candidate sprint learning paths to obtain a plurality of iteratively optimized sprint learning paths; in any iterative optimization, if the current candidate sprint learning path violates a first repair relation or / and a time window or / and a daily learning duration budget in the learning path, performing feasibility repair on the current candidate sprint learning path; performing neighborhood local search on the iteratively optimized sprint learning path based on a forgetting risk index; and outputting an optimal sprint learning path.

The invention relates to an attachment certificate dynamic form processing method and system, and belongs to the technical field of enterprise financial certificate intelligent processing, and the method comprises the steps: receiving an attachment certificate uploaded by a user, and extracting metadata features such as a file size, a format type and a hash value to generate an attachment object; carrying out fragmentation storage on the object by adopting an Erase Coding algorithm, and generating a voucher type identification cue word based on metadata; analyzing the voucher through a hybrid recognition model fusing a convolutional neural network and a Transform encoder, and outputting structured data containing a voucher type identifier; mapping the data into a database entity and triggering a voucher life cycle state change event, and driving a business rule engine to dynamically generate a form mode definition associated with a voucher type; and in response to user query, generating an interactive form through a responsive rendering engine in combination with a voucher type, state and authority dynamic combination form mode and instance data.

The invention discloses a high-speed fixed-point multiplication circuit. The multiplication circuit is mainly composed of a multiplier coding module, a partial product generation module, a partial product compression module and a traveling wave carry adder module. The multiplier coding module is composed of a radix-4-Booth coding algorithm and an opposite number generation module, and is used for carrying out three-bit block coding on input multiplication data and generating an opposite number of a multiplicand in advance. The partial product generation module generates a plurality of groups of partial products with symbol extension according to the multiplicand coded signal. And the partial product compression module adopts an improved Wallace compression structure to perform layered compression on the partial product. And the traveling wave carry adder module sums the two groups of results output by compression and outputs a multiplication result. According to the invention, multiplication accumulation series can be reduced, the switching times of invalid signals in the circuit can be reduced, and the longest delay path in an operation link can be shortened, so that fixed-point multiplication which is high in speed, low in power consumption and more favorable for a comprehensive tool in structure is realized.

The invention discloses an artificial intelligence model training method and system based on a block chain, and relates to the technical field of block chains, and the method comprises the steps: collecting physical position coordinate data, and converting the data into a position coding character string through a position coding algorithm; generating an adjacent target training node list according to a prefix matching rule of the position coding character string, and triggering the target training nodes through block chain transaction to create a local training group; the target training node initiates real-time training gradient exchange in the local training group; when the exchange fails, requesting historical gradient data associated with the position coded character string from a gradient bank; the artificial intelligence model is updated by using historical gradient data, and a new training gradient is generated after updating is completed; and storing the new training gradient to a partition position corresponding to the position coding character string. According to the method, the artificial intelligence model is updated through the historical gradient data associated with the position coding character string to generate the new training gradient, and self-recovery of model iteration in a communication interruption scene is realized.

The application discloses a fine-tuning fault-tolerant method, device, medium and product of an intelligent question and answer model, relates to the technical field of artificial intelligence, and comprises the following steps: dividing a preset low-rank matrix pre-injected in an intelligent question and answer model into at least two target sub-matrices according to the influence degree of the matrix parameters on the model performance, and matching corresponding target coding algorithms for different target sub-matrices, so that hierarchical and differentiated data protection is realized, and resource waste caused by uniform intensity protection on all parameters is avoided. Moreover, the target sub-matrices are respectively subjected to coding calculation at a first preset time point and a second preset time point of any two adjacent training steps, and the two coding results are compared, so that a data damage event occurring to the preset low-rank matrix in a training process can be effectively and quickly detected, thereby preventing error pollution from spreading to the whole model, and the reliability of the system is improved. The scheme disclosed by the application can significantly reduce fault-tolerant overhead and guarantee the stability in the fine-tuning process of the model.

The invention provides a lightweight storage method and system for time series data of the Internet of Things, and relates to the technical field of data storage and compression of the Internet of Things, and the method comprises the steps: collecting multi-mode time series data and mechanical vibration energy data generated during the operation of equipment of the Internet of Things; extracting vibration frequency characteristics from the mechanical vibration energy data, and generating a time sequence vibration signal based on the vibration frequency characteristics; performing space-time correlation on the time sequence vibration signal and the multi-modal time sequence data to identify a data activeness level; according to the data activeness level, performing sparse representation processing on the multi-modal time series data to obtain sparse representation data; and carrying out dynamic dictionary construction and coding on the sparse representation data by adopting a self-adaptive dictionary coding algorithm to generate compressed coding data, and carrying out hierarchical storage management on the compressed coding data to realize lightweight storage. According to the invention, the storage efficiency and the processing capability of the time series data of the Internet of Things edge device are improved.

The technology disclosed relates to a system for identifying same or similar instances of source code. In particular, the technology disclosed relates to a data repository designed for storing information regarding one or more associations between an instance of source code and a semantic hash code value. Wherein, semantic hash code value being determined in accordance with a semantic hash code algorithm. Wherein, semantic hash code algorithm being designed for computing a quantification of a set of measured characteristics of instance of source code. Wherein, quantification of set of source code characteristics of instance of source code being represented by semantic hash code value. Wherein, semantic hash code value corresponding to instance of source code. Wherein, an association between an identity of instance of source code and semantic hash code value is stored into data repository.

The application provides a time window-based time series data compression hierarchical storage method and system, relates to the technical field of time windows, and the method comprises the following steps: firstly, collecting the running time series data of an industrial device, dividing the time series data into data blocks according to fixed time windows, and extracting statistical features of each data block to generate working condition feature data; then, analyzing the energy distribution features of the data blocks, generating optical pulse coding sequences through photoelectric pulse conversion and pulse width modulation technology; determining the compression level through analysis of the pulse density distribution, and combining the working condition feature data to perform compression processing by using a corresponding coding algorithm; finally, distributing the data to different storage areas according to the compression level to complete hierarchical storage. The application improves the storage efficiency of industrial monitoring data and reduces the storage cost.

Embodiments of the present disclosure provide a machine video feature encoding method, a decoding method, an apparatus, a device, a medium and a program product. The method converts a machine video feature group into a feature tensor graph group, and respectively performs two rounds of rearrangement on the feature tensor graph and each feature tensor in the feature tensor graph to generate a second feature graph. Meanwhile, the position change information of the rearrangement is recorded through a serial number matrix. Based on the second feature graph and the serial number matrix, a compressed code stream is obtained. Through the rearrangement process, the generated second feature graph is suitable for encoding and decoding by using a video coding algorithm, thereby saving bandwidth, improving transmission speed and reducing the cost of related processing without changing the existing video codec hardware system.

The application relates to a multiplier based on radix-4 Booth coding and an improved Wallace compression tree, comprising a partial product generation module: input multipliers are re-encoded into a plurality of partial products by adopting a radix-4 Booth algorithm to form a compact partial product array; a partial product compression module: by utilizing an improved Wallace tree compression structure, the partial products are quickly compressed by two compressors to reduce the calculation path delay; and a final summation module: the compressed partial products are summed by using a high-speed adder to obtain a final multiplication result. The radix-4 Booth coding algorithm is adopted to make the partial product array arrangement regular, the circuit area is reduced, a new tree compression structure in which 3-2 compressors and 4-2 compressors are mixed is adopted to realize the Wallace tree, the compression efficiency is improved, and the key delay is reduced.

The application relates to a video coding method, device, computer equipment and storage medium. The method comprises the following steps: obtaining a region of interest and a non-region of interest in a current video frame of a video stream; using a first coding strategy to code the region of interest, using a second coding strategy to code the non-region of interest, and outputting coding data; wherein the coding quality of the first coding strategy is higher than that of the second coding strategy. The application can effectively control the code rate fluctuation by optimizing the coding algorithm to perform smooth processing on the data packet, thereby reducing the jitter and congestion risk in network transmission, and especially improving the picture freezing problem caused by large I frame transmission delay in the real-time video transmission scene.

This invention relates to the field of computer software technology. Specifically, it relates to a method for constructing a software development model adapted to multiple hardware platforms. When the target hardware platform is first started, a hardware baseline detection daemon is run to collect multi-dimensional low-level physical feature parameters and uses a sparse coding algorithm to fuse and generate a hardware baseline capability fingerprint. During the software development phase, a compile-time annotation processor is used to scan key functional modules and extract non-functional feature contract tags pre-embedded by the developer using structured declarative annotations into a tag list index file. After the fingerprint generation is completed upon software startup, a multi-dimensional constraint solving rule engine is used to perform conflict detection and optimal fit scoring and sorting of the module contract tags and hardware fingerprints, and to select the uniquely matching target functional module entry class. Based on this entry class, a server-side dynamic orchestration and distribution mechanism is used to extract only the associated high-cohesion atomized payload packages and inject them into the main program runtime context to complete dynamic assembly.

The invention relates to the technical field of integrated circuit design, in particular to a processing circuit architecture supporting multi-calculation precision dynamic switching, which comprises a control port module, a partial product generation module, a symbol compression module, an addition compression tree module and a final adder module. And single-cycle dynamic switching of various precisions is realized. The partial product generation module adopts a Booth coding algorithm to split an operation vector, and cooperates with boundary symbol selection logic to solve the problem of symbol expansion and auxiliary bit overlapping; the symbol compression module compresses the redundant extension bits through a preset coding mode; the addition compression tree module is formed by cascading multiple stages of compressors and inserting carry blocking logic; the final adder module is composed of a plurality of carry lookahead adders, and the output bit width is dynamically controlled through blocking logic. The architecture optimizes the parallel operation efficiency and the resource utilization rate, adapts to the deep learning training and reasoning full scene, and has the advantages of real-time performance and low power consumption.

The invention discloses a multi-prover zero-knowledge proving method without prover communication, relates to the technical field of multi-prover zero-knowledge proving methods, solves the defect that provers need frequent communication in the multi-prover zero-knowledge proving method in the prior art, and comprises the following steps of: setting an algorithm: inputting a safety parameter lambda and a proposition P to be proved, the verifier outputs N certification keys {EK1,..., EKN} and a verification public key VKF; according to the coding algorithm, a proposition input u to be proved is input, and a verifier outputs a code Iu; according to the certification algorithm, a server index b and a bth certification key EKb are input, a code Iu is input, and a bth server outputs a bth proposition result and a bth certification pi (y, b); the verification algorithm comprises the steps of inputting a verification public key VKF, proposition input u, N proposition results and N certificates {pi (y, b)} (b = 1) N, if verification succeeds, outputting 1, and otherwise, outputting 0. The decentralized cooperation generation proof thoroughly solves the problem of communication dependence.

The invention discloses a GPU memory optimization method and system, and the method comprises the steps: obtaining the real-time change characteristics of a task load through obtaining the dynamic data flow information of the task load and an intermediate data generation rule; according to the real-time change characteristics of the task load, an optimization algorithm based on gradient descent is adopted to determine the distribution proportion and the reserved space of a memory pool, and adjusted memory pool configuration is obtained; extracting available memory block information from the adjusted memory pool configuration, and compressing the intermediate data with low access frequency by adopting a Huffman coding algorithm to obtain a compressed intermediate data storage unit; and if the access frequency of the compressed intermediate data storage unit is lower than a preset threshold value, temporarily storing the compressed intermediate data storage unit to a low-speed storage area, and reallocating memory resources according to the task priority to obtain an optimized memory allocation scheme. According to the method, the memory utilization rate is remarkably improved, the access delay is reduced, and efficient memory resource management is realized.

Embodiments of the present disclosure provide a watermark embedding method, a watermark extraction method and device. The watermark embedding method comprises the following steps: first, obtaining a to-be-processed field in a to-be-processed text and a watermark insertion length of the to-be-processed field; then, encoding identification information into watermark code based on a coding system of a homograph code table and a digital coding algorithm; determining a to-be-encrypted watermark corresponding to the watermark code based on the watermark insertion length and the watermark code; encrypting the to-be-encrypted watermark based on a key to obtain an encrypted watermark; and finally, encoding the encrypted watermark into the to-be-processed field based on the homograph code table to generate a target field with watermark code, wherein the watermark is embedded by using a control character of the homograph code table on the basis of a general webpage watermark, and only the underlying coding and display form of the to-be-processed field are changed, without any change to the text format and expression content, thereby improving the security of the to-be-processed text and the target field in the to-be-processed text.

The application discloses a compression method and device of seismic data, electronic equipment, storage medium and product. The method comprises the following steps: extracting pre-stack seismic data into common offset data; performing discrete wavelet transform on the common offset data to obtain transformed data; performing multi-level coding on the transformed data based on a multi-level tree set splitting coding algorithm to obtain compressed data, wherein the coding level number of the multi-level coding includes an integer from a high coding level number to a maximum coding level number, the maximum coding level number is the total number of thresholds that can be used to determine the importance of wavelet coefficients, and the high coding level number is an integer greater than 1 and less than the maximum coding level number. The above scheme realizes seismic data compression by performing multi-level coding on the wavelet domain data of the seismic data through the multi-level tree set splitting coding algorithm, can guarantee the compression quality of the seismic data, and starts positioning and coding of the data from the high coding level number, so that the total coding level number is reduced, and the compression efficiency of the seismic data is improved.

This invention proposes an unconditional image generation method and system based on deep predictive coding networks, comprising: step S1, optimizing the network structure to implement a predictive coding algorithm with skip connections; step S2, building a training framework for the co-evolution of the input and output ends based on the original predictive coding algorithm; and step S3, improving the original predictive coding algorithm by introducing a non-convex energy function optimization algorithm into the predictive coding algorithm. This invention avoids gradient vanishing, accelerates training speed, and optimizes generation quality. Deep backpropagation (BP) convolutional networks suffer from gradient reduction, a problem first addressed in ResNet with skip connections. Skip connections directly send information from lower layers closer to the input to higher layers closer to the output, facilitating rapid gradient calculation to support parameter updates. This invention achieves unconditional image generation based on the original predictive coding and improves the stability of numerical computation.

The invention relates to an aviation structure impact load non-negative sparse coding algorithm expansion identification method and system. The method comprises the following steps: acquiring a monitored aviation structure and making a data set comprising an impact load force signal and a response signal; constructing a non-negative sparse coding model based on an impact load dynamic model and non-negative sparse prior; constructing an iterative convergence threshold algorithm for solving the non-negative sparse coding model; constructing a depth algorithm expansion network based on the iterative convergence threshold algorithm; training the depth algorithm expansion network by using the data set to obtain a network model with optimal parameters; and performing an impact load identification test by using the network model with the optimal parameters, and outputting an impact load to be identified. The method has the advantages that the calculation speed is high, parameters do not need to be manually set, and good noise immunity is shown; the neural network provided by the invention has better interpretability in structural design.

The invention discloses a large model Token coding analysis method and device, equipment and a medium, and relates to the technical field of natural language processing, and the method comprises the steps: determining a task demand after receiving a target natural language processing task, and reading text data in a preset training corpus according to a target coding format; processing the text data based on task requirements by utilizing a byte pair coding algorithm to obtain a target vocabulary; determining a to-be-processed text, and performing hierarchical coding on the to-be-processed text based on the target vocabulary by using a preset hierarchical coding strategy to obtain a target Token sequence; and inputting the target Token sequence into a preset large model, so that the preset large model performs semantic analysis on the target Token sequence to obtain a corresponding semantic analysis result and an updated Token sequence, and performing analysis processing on the updated Token sequence based on the semantic analysis result to obtain an analyzed Token sequence.

The application provides a unified learning method and system supporting balanced propagation and predictive coding, and belongs to the technical field of neural network calculation and optimization. The method comprises the following steps: determining the network structure and connection type of a unified learning model supporting a balanced propagation algorithm and a predictive coding algorithm based on the calculation graph of the balanced propagation algorithm and the predictive coding algorithm; extracting the energy function of the balanced propagation algorithm and the predictive coding algorithm, configuring the interaction term of the energy function, and taking the configured energy function as the energy function of the unified learning model to support multi-convolution network calculation; and extracting the network structure, connection type and energy function of the unified learning model at a user end to construct a qualified multi-layer neural network structure, and optimizing the multi-layer neural network structure by using an optimizer.

The invention relates to a trusted transmission method and equipment for communication data of power equipment and a medium, and belongs to the technical field of data processing, and the method comprises the following steps: obtaining communication data to be transmitted and an equipment identifier of the power equipment, and packaging the communication data into protection data by adopting an HMAC algorithm based on the equipment identifier; taking the protection data as the input of a Reed-Solomon coding algorithm, and outputting coded data; taking the coded data as the input of an AES-CTR encryption algorithm, and outputting encrypted data; performing time signature on the encrypted data by adopting an RSA-PSS signature algorithm to obtain signature data; and obtaining identity binding data based on the signature data, the encrypted data and the device identifier, and sending the identity binding data to a receiver. According to the invention, through deep binding of the equipment identifier, the encrypted data and the signature data, accurate identity authentication and traceability functions of the communication data of the power equipment are realized.

The invention discloses a front-end development optimization method and system, and the method comprises the steps: generating a unique dynamic routing identifier through a dynamic feature coding algorithm according to the operation type of a user and the features of terminal equipment; generating a service-route real-time binding table based on the dynamic route identifier; generating a multi-terminal adaptive atomization rendering component set through a syntax tree conversion engine according to the rendering specification difference of the cross-terminal platform; generating a cross-platform incremental update instruction set based on the real-time content update request; and executing the incremental updating instruction set by adopting a resource loading optimization strategy according to terminal operation environment performance parameters, eliminating cross-end compatibility conflicts through a rendering pipeline recombination technology, and outputting a multi-terminal consistent content presentation effect. By utilizing the embodiment of the invention, the key problems of multi-terminal adaptation, performance optimization, real-time updating and the like can be systematically solved, and the front-end development efficiency is improved.

The invention provides a time window-based time sequence data compression hierarchical storage method and system, and relates to the technical field of time windows.The method comprises the steps of firstly, collecting operation time sequence data of industrial equipment, dividing the operation time sequence data into data blocks according to a fixed time window, and extracting statistical characteristics of each data block to generate working condition characteristic data; then analyzing the energy distribution characteristics of the data blocks, and generating an optical pulse coding sequence through photoelectric pulse conversion and pulse width modulation technologies; a compression level is determined through analysis of pulse density distribution, and compression processing is performed by adopting a corresponding coding algorithm in combination with working condition characteristic data; and finally, distributing the data to different storage areas according to the compression level to complete hierarchical storage. The industrial monitoring data storage efficiency is improved, and the storage cost is reduced.

The application discloses a sprint learning path optimization method based on an improved beluga whale optimization algorithm, and solves the problem that existing sprint path planning is unreliable. The method comprises the following steps: constructing a total learning cost target function; adopting a random key coding algorithm based on chaotic mapping to represent a learning path in a sequence manner, and adopting serial scheduling to obtain a plurality of candidate sprint learning paths; introducing an adaptive weight mechanism, an examination-driven offset term and a Gaussian disturbance term, and adopting an improved beluga whale optimization algorithm to iteratively optimize the candidate sprint learning paths, so as to obtain a plurality of iteratively optimized sprint learning paths; in any iteration optimization, if the current candidate sprint learning path violates a prerequisite relationship in the learning path, or / and a time window, or / and a daily learning time length budget, then the current candidate sprint learning path is repaired for feasibility; the iteratively optimized sprint learning paths are subjected to neighborhood local search based on a forgetting risk index; and an optimal sprint learning path is output.

The application discloses a semantic tendency analysis system based on dynamic word segmentation and multi-round inquiry, comprising the following modules: a character-level embedding construction module, which is used for performing embedding coding and generating a word embedding sequence; a multi-granularity context coding module, which is based on a multi-granularity context coding algorithm and outputs a multi-granularity context feature set; a semantic fusion module, which is used for calculating the difference degree of the multi-granularity context feature set and learning the fusion weight, and generating a fused semantic vector sequence; a neighboring word boundary perception module, which is used for constructing a boundary judgment node, generating a boundary judgment feature vector, and mapping the boundary judgment feature vector into a boundary confidence sequence; a dynamic word segmentation module, which is used for constructing a dynamic threshold function, calling a multi-round inquiry mechanism, and outputting a structured word sequence; and a semantic tendency discrimination module, which is used for performing word vector mapping and context modeling, and outputting a sentiment label result and a corresponding confidence score. The application constructs a high-precision semantic tendency analysis system suitable for a complex context.

This invention discloses a GPU memory optimization method and system. By acquiring dynamic data flow information and intermediate data generation patterns of the task load, the real-time variation characteristics of the task load are obtained. Based on these characteristics, a gradient descent-based optimization algorithm is used to determine the memory pool allocation ratio and reserved space, resulting in an adjusted memory pool configuration. Available memory block information is extracted from the adjusted configuration, and a Huffman coding algorithm is used to compress low-access-frequency intermediate data, resulting in compressed intermediate data storage units. If the access frequency of the compressed intermediate data storage units is lower than a preset threshold, the compressed intermediate data storage units are temporarily stored in a low-speed storage area, and memory resources are reallocated according to task priority, resulting in an optimized memory allocation scheme. This invention significantly improves memory utilization, reduces access latency, and achieves efficient memory resource management.

The invention relates to an unmanned carry-scraper video transmission method based on RK3588S and AHD cameras, and belongs to the technical field of video transmission. The method comprises the following steps: acquiring image data in real time by adopting a plurality of AHD cameras so as to obtain an unmanned carry-scraper video; the AHD camera has the functions of automatically adjusting exposure, automatically adjusting white balance and automatically supplementing light; according to the method, an RK3588S development board is utilized, H.265 hard coding is adopted when a video is coded, coding parameters are dynamically adjusted, and a coding algorithm is optimized; in a wireless environment, a multi-path transmission technology is adopted to carry out video transmission; and decoding and displaying the transmitted video. Compared with the existing unmanned carry-scraper video transmission technology, the method has the advantages of high efficiency, low time delay and low CPU (Central Processing Unit) occupation, the time delay of a video display end is reduced from 2 seconds to only 0.3 seconds, and the CPU occupation is reduced from 500 percent to 100 percent.

The application discloses a multi-dimensional time series data lossy compression method based on dimension correlation grouping. The method extracts dynamic trend characteristics of each dimension data, calculates dynamic error boundary of each dimension, constructs high correlation dimension grouping and determines primary and secondary dimensions in each grouping; introduces error feedback mechanism in time domain to perform differential processing on the data, generates quantized differences of each dimension data; calculates differential residuals of secondary dimensions in each group relative to the primary dimension, adjusts compression strategy according to the amplitude characteristics of the residuals, and generates a to-be-encoded set composed of a primary dimension sequence and a secondary dimension residual sequence; and performs compression on the to-be-encoded set through an entropy coding algorithm and encapsulates the to-be-encoded set. The compression strategy of the application adopts correlation grouping between dimensions, feedback quantization, residual construction between dimensions in a group and bit reorganization, and significantly improves compression density and reconstruction accuracy for large-scale sensor array data.

The invention discloses a method for optimizing the transmission quality of a mobile device. The method comprises the following steps: S1, acquiring transmission quality parameters of each channel and preprocessing an original data stream to be transmitted; s2, dividing the preprocessed data stream into k data packets by adopting a Reed-Solomon coding algorithm, and generating r redundant data packets; s3, adding a serial number, a timestamp and channel identification data to each data packet; s4, intelligently distributing the data packets to different channels based on the transmission quality parameters; s5, detecting the packet loss condition of the channel, and if continuous packet loss occurs, trying to reconstruct data by using an FEC (Forward Error Correction) mechanism; s6, if data reconstruction fails, dynamically adjusting a data distribution strategy; s7, performing FEC decoding on the received data at the receiving end; and S8, recovering the successfully decoded data into a complete original data stream. According to the method, the transmission reliability and real-time performance in a complex network environment are remarkably improved, and the method is suitable for a high-reliability data communication scene.