63 results about "Computational algorithm" patented technology

The invention relates to the technical field of production scheduling, and discloses a production scheduling visualization method based on big data. According to the method, multi-dimensional production data such as equipment real-time states, material inventory, personnel skills and order delivery periods are collected, and a four-dimensional dynamic constraint model is generated through a dynamic constraint modeling algorithm; and obtaining an initial scheduling scheme by using a mixed integer programming and improved genetic algorithm dual-drive optimization engine, inputting the initial scheduling scheme into a visual interaction module, displaying a result by using a drag-and-drop Gantt chart, and detecting resource conflicts in real time. The feasibility of the adjusted scheduling scheme is verified through a dynamic re-calculation algorithm, and then the model is updated. According to the method, multi-dimensional data are integrated, production scheduling optimization and visualization are realized, production resource conflicts can be effectively solved, the production efficiency is improved, the cost is reduced, and an efficient solution is provided for enterprise production management.

The invention provides a house waterproof maintenance system based on an intelligent detection and self-repairing technology, and the system comprises an intelligent sensing layer which is used for collecting multi-dimensional physical parameters of building waterproof key nodes in real time; the edge calculation analysis layer is used for performing fusion analysis and risk prediction on the perception data; the dynamic self-repairing execution layer is used for triggering a multi-stage repairing mechanism according to an analysis result; the cloud management layer is used for realizing visual management and control and strategy optimization of the whole system, and relates to the technical field of building waterproofing, the building state is monitored in real time through a multi-mode sensing network, water leakage risk prediction is realized in combination with an edge calculation algorithm, and collaborative repair of a microcapsule repair agent, a shape memory alloy grid and an electrodialysis module is triggered; the system remarkably improves the timeliness and reliability of waterproof maintenance, reduces the manual intervention cost, and is suitable for intelligent maintenance of residential buildings, industrial buildings and historical protection buildings.

The invention discloses an image processing method and device, equipment and a storage medium, and is applied to the technical field of image processing. According to the scheme, the method comprises the following steps: detecting angular points in a target image to be interpolated based on an angular point detection algorithm to obtain feature points in the target image; for the texture complexity of the neighborhood image of each feature point, selecting a target interpolation algorithm for processing the feature point; and performing interpolation processing on the feature points based on a target interpolation algorithm to generate an interpolation result image. For angular points with complex textures, a high-precision interpolation algorithm is adopted to ensure that image details are finely restored; and for the angular points with slightly simple textures, a low-calculation-amount algorithm is used to reduce resource occupation, so that the limitation that a single algorithm is uniformly adopted for the whole image in the prior art is avoided, and the redundant calculation amount is remarkably reduced while the interpolation precision of the key feature region is ensured and the quality of the whole image is maintained.

The invention provides a training and examination robot system based on a multi-modal large model, equipment and a computer program product, and aims to accelerate training transformation and upgrading and improve coping ability as basic appeals, and based on regular training work of teams, training and examination robot, sensor acquisition equipment and big data platform three-terminal equipment. A multi-modal data acquisition layer for realizing real-time attitude detection and motion capture by using an edge calculation algorithm, a data processing and analysis layer for finishing real-time image processing through localized edge calculation in a network-free or power-free environment, and system control functions for finishing intelligent decision making, evaluation, error correction and the like are provided; an intelligent decision-making and control layer of a training subject rewriting decision-making mechanism is established, cross-modal collaborative learning middleware is utilized to realize modal representation consistency while training subject transposition evaluation is realized, training and assessment operations of the same subject of multiple persons or different subjects of multiple persons are supported, and the training and assessment efficiency is remarkably improved.

Described herein are digital twin methods and systems integrating advanced computational algorithms and user-centric interfaces to transform the management of gas turbine engines, which provide comprehensive real-time analytics, extending beyond basic performance prediction, design, analysis and monitoring as well as capabilities for predictive maintenance, design and efficiency optimization, and fault diagnosis, all accessible to operators regardless of their expertise in turbine physics.

Computational algorithms or “solvers” may be executed on intelligent electronic devices (IEDs) in an electric power delivery system to perform tasks such as locating and isolating electrical faults and restoring service to an area after an electrical fault occurs. However, executing the solvers may become a computationally prohibitive task, particularly where computing power is limited (e.g., low-power pole-top computers). In certain embodiments, computing power may be offloaded onto multi-instance networks that may enable processing of large amounts of complex data. It may be difficult for the multi-instance networks to obtain real-time or near real-time data from the electric power delivery system due to cybersecurity concerns. As such, multiple stateless services may be used to evaluate limited and / or obfuscated representations of a portion (e.g., a single feeder circuit) of the electric power delivery system.

The present invention discloses an online recommendation method for remote sensing image processing algorithms, which belongs to the field of remote sensing image processing. The evaluation method includes: constructing a standard data set required for algorithm evaluation; receiving a remote sensing image processing algorithm and a description file uploaded by a user; parsing the description file of the remote sensing image processing algorithm; matching a set of corresponding standard input data according to the algorithm type, using each standard input data as an input parameter to drive the remote sensing image processing algorithm to execute once; recording resource usage information, algorithm execution time, and algorithm output parameters, and calculating the algorithm accuracy; performing weighted summation on the resource usage information, algorithm execution time, and algorithm accuracy obtained from previous algorithm executions to obtain a comprehensive score for the algorithm. Based on the evaluation, the recommendation method constructs an algorithm optimization deep learning model and recommends the most suitable algorithm to improve image processing efficiency.

The invention provides a full-slice image quality automatic control method and system based on a two-stage cascade deep neural network, and relates to the technical field of digital pathology image processing. According to the method, efficient, accurate and extensible automatic quality control is realized through the steps of a two-stage cascade detection architecture, a multi-scale feature fusion mechanism, standardized GeoJSON output and multi-dimensional quality scoring. The core of the method is that priori knowledge of pathology (firstly positioning organization and then checking quality) is converted into strict conditional probability decomposition, and the method is efficiently realized through a deep learning network. The mapping from the human cognitive process to the computational algorithm provides a new methodological enlightenment for the field of medical image analysis. Theoretical analysis and actual tests show that the method is remarkably superior to the prior art in the aspects of calculation efficiency, detection precision, cross-device generalization ability and the like, and has important clinical application value and commercial prospects.

The present invention discloses a real-time data collection and processing method and system for a smart park based on edge computing, which relates to the field of data collection technology and is used to improve the problem of different display effects caused by the use of a unified processing method for park images at different locations. The method comprises obtaining attribute data and point information of each collection device in the park, setting a collection force scoring mechanism to score the collection force of different collection devices using fuzzy reasoning, dividing the park into regions according to the collection force score and point information, performing real-time detection on the detection area and the number of faults of the collection equipment in each divided area and calculating the real-time coverage rate of each divided area using a polynomial regression algorithm, classifying each divided area, obtaining regional images of low real-time coverage areas for marking and collecting the amount of dust and humidity in the space, calculating the humidity transfer rate, and judging whether the corresponding area uses an enhanced edge computing algorithm to process the regional image by using a geometric mean method.

Systems and methods for utilizing imperfect resource states in a fusion based quantum computing system. First and second entangled quantum states are received and input to a quantum computational algorithm that performs a fusion measurement on the first and second entangled quantum states. It is determined that one of the first or second entangled quantum states have an error that exceeds an error threshold. Responsive to the error exceeding the error threshold, the quantum computational algorithm is dynamically modified such that an alternative process resulting in classical measurement results is performed instead of the fusion measurement. The classical measurement results are stored in a non-transitory memory medium.

The invention relates to the technical field of audio processing, in particular to an audio gain output dynamic adjustment method and device, equipment, a storage medium and a computer program product. The method comprises the following steps: establishing a corresponding volume sequence based on initial audio data, and storing the volume sequence in a local database; based on the volume operation instruction of the user, updating the gain parameter, and recording boundary information corresponding to the volume operation instruction in a local database; determining a target playing range according to the boundary information, and generating a mapping relation between the input volume and the output volume; based on the mapping relation, processing the initial audio data by adopting a preset edge calculation algorithm to obtain an adjusted audio frame; and controlling the audio and video playing device to output the adjusted audio frame according to the updated gain parameter, thereby improving the audio playing quality of the audio and video playing device.

The disclosure is directed to a method including executing, at a first time, a precompute algorithm that is a first portion of a quantum algorithm. Executing the precompute algorithm generates a precompute output that includes first quantum information encoded in a first set of qubits. A runtime input for the quantum algorithm is received at a second time that is subsequent to the first time. A runtime algorithm is executed, at a third time that is subsequent to the second time. The runtime algorithm is a second portion of the quantum algorithm. Executing the runtime algorithm is based on the first quantum information encoded in the first set of qubits. Executing the runtime algorithm generates a runtime output that includes second quantum information encoded in a set of qubits. An output of the quantum algorithm is provided. The output of the quantum algorithm is based on the second quantum information.

A first method includes capturing, via a camera having a first pose, a first image of a surgical environment after a change has been made to the surgical environment. The method also includes capturing, via. a. camera having a second pose that is different from the first pose, a second image of the surgical environment after the change has been made to the surgical environment. Hie method also includes using a computational algorithm to process the first image, the second image, and a model representing the surgical environment, thereby updating the model to represent the change made to the surgical environment. A second method includes training the computational algorithm to perform aspects of the first method.

Disclosed is a unified deep reinforcement learning approach for constructing rectilinear and octilinear Steiner minimum trees in the technical field of computer-aided design of integrated circuits, the method including designing an edge point sequence (EPS) based on structural characteristics of a Steiner minimum tree (SMT) to bridge a gap between deep learning model output and an SMT structure; designing a deep learning model for the EPS, and using a negative wirelength of the SMT as a reward to train the deep learning model through the deep reinforcement learning (DRL); providing a corresponding fast and accurate wirelength computation algorithm for a quality assessment of construction solutions to accelerate training of the deep learning model; and constructing diversified construction solutions of SMTs utilizing the stochastic nature of machine learning. The method of the present invention can solve rectilinear Steiner minimum tree (RSMT) and octilinear Steiner minimum tree (OSMT) problems and generate diversified routing topologies.

The invention discloses a high-order extended Kalman filtering method for a nonlinear system. The method comprises the following steps: S1, establishing a nonlinear system composed of a linear state model equation and a nonlinear measurement model equation; s2, establishing a calculation algorithm for solving the statistical characteristics of the Kronecker product type composite random variables; s3, analyzing the high-order characteristics of the state estimation error of the nonlinear system by using the algorithm established in the step S2; s4, combining the known measurement information with the high-order features of the errors, designing and optimizing extended Kalman filtering, and gradually estimating the next state in the nonlinear system; and S5, circulating the process of the step S4, estimating the state of the nonlinear system step by step until the state estimation of the set step number is completed, and obtaining the state estimation value of each step. According to the method, the estimation model in the filtering process is adaptively adjusted based on the high-order statistical characteristics of the system state estimation error and the known measurement information, so that the stability and the precision of the filter are improved.

The invention discloses a method for testing the performance of a neural network acceleration chip in a multi-task parallel scene, and the method comprises the steps: setting nine test indexes according to an object loaded by the neural network acceleration chip, and completing the test of the task operation performance of the neural network acceleration chip in the multi-task parallel task scene through the steps 1-9. And evaluating the task switching capability in different scenes, different entities and different tasks on the basis. The task scene data generation is responsible for generating a task environment and sensor data in a task execution process. And driving a related neural network algorithm to complete the task according to the task content and requirements. In the algorithm evaluation process, the operation information of the algorithm is collected, and the evaluation index of the algorithm is calculated according to the overall effect of task execution, so that the multi-task parallel capability of the neural network acceleration chip is measured.

The invention provides a method for scheduling a test task, and relates to the technical field of computers, and the implementation scheme is as follows: obtaining a to-be-scheduled test task, a test data set and the state of computing power and bandwidth resources; based on the test task, the states of the test data set and the computing power and bandwidth resources, calculating an algorithm feature vector, a data set feature vector and a test strategy feature vector; constructing a ternary relation graph based on the algorithm feature vector, the data set feature vector and the test strategy feature vector; performing feasibility judgment and cost evaluation on each ternary relation graph; based on feasibility judgment and cost evaluation, selecting a target ternary relation graph to generate a task execution scheme; and according to the task execution scheme, binding the test task to the matched resource slot and data set for execution, so that algorithm features, data set features and test strategies can be brought into a decision together, coverage of key scenes can be completed preferentially, the resource utilization rate is improved, and queuing delay and data handling cost are reduced.

The one or more embodiments of the application provide a method, device and storage medium for evaluating running of a privacy computing algorithm, the method comprising: obtaining algorithm running information corresponding to a privacy computing algorithm to be evaluated and device configuration information of a device for running the privacy computing algorithm; inputting the algorithm running information and the device configuration information into a service model of a large language model, to predict, by the service model of the large language model, a running condition of the privacy computing algorithm on the device based on the algorithm running information and the device configuration information; the service model of the large language model being a model obtained by performing supervised fine-tuning training on a base model of the large language model based on training samples including the algorithm running information and the device configuration information and labeled with the running condition; obtaining the running condition of the privacy computing algorithm on the device predicted by the service model of the large language model, and outputting the running condition.

The application discloses a multiplexing measurement and control method and device of a three-axis micromechanical gyroscope, which comprises the following steps: firstly, a global clock is generated; in the working clock of any gyroscope, an alternating current driving signal, a feedback force signal, a preset reference phase signal and a reference displacement signal are outputted to the gyroscope by a coordinate rotation digital computer, and the gyroscope reaches a force balance state through a driving modal control loop and a detection modal control loop, and the angular velocity of the gyroscope is obtained; the three gyroscopes multiplex the same coordinate rotation digital computer in the respective working clock. The application combines time division multiplexing with a coordinate rotation digital computer algorithm, and can accurately and synchronously detect and close-loop control the three-axis micromechanical gyroscope in real time in limited field programmable logic gate array resources. The method and device can be widely applied to concurrent processing requirements of multi-axis or multi-path high-speed, high-precision and high-integration measurement and control instruments.

Disclosed is a machine learning method for curating biomarker panel, for screening of cancer(s) and system therefor, addressing the critical need for early detection through a combination of innovative biomarker research, advanced computational algorithms and integration with existing healthcare infrastructure. The biomarker panel is curated to facilitate pan-cancer screening of the patient with high sensitivity and specificity using the ML method deploying MoE (Mixture of Expert) models for accurate screening, thereby enhancing diagnostic accuracy and also empowering clinicians with actionable insights, thereby improving patient prognosis and facilitating personalized treatment strategies.

The application discloses a hardware-implemented H264 macroblock-level code rate control method, which comprises the following steps: using a five-stage pipeline to implement H264 hardware coding, specifically, IME module, FME module, IPTQ module, DBK module, EE module and WBK module are used; and the hardware-implemented macroblock-level code rate control method mainly comprises the following steps: when calculating a macroblock quantization parameter QP, the calculation is divided into two stages, coarse quantization parameter coarse_qp is calculated through texture information of the macroblock, and the value is only used for the IME module and the FME module; after the macroblock type of the macroblock is determined, the mad value of the macroblock can be obtained, and the macroblock quantization parameter mb_qp is recalculated according to the mad value of the macroblock, and the value is used for operation and processing of the IPTQ module, the DBK module and the EE module. The method avoids complex macroblock quantization parameter QP calculation in the JVT-G012 algorithm, and the algorithm is simple and easy to implement in hardware.

The embodiment of the invention provides a real-time active feedback sleep monitoring system and method based on a ring. According to the system and the method, sound information data and physiological information data acquired by a ring are acquired through a microphone array of the charging dock, localized real-time analysis of the ring is realized by adopting an edge calculation algorithm, and multi-stage vibration tactile feedback is generated based on a vibration motor arranged in the ring, so that real-time active feedback can be realized during risk pre-judgment.

The application provides a high-voltage bushing discharge sensitivity test method simulating dust invasion influence, belongs to the technical field of high-voltage bushings, and is based on a partition asynchronous propulsion algorithm to solve a charge density gradient field and a dust motion trajectory field, carries out wavelet packet multi-scale decomposition on a partial discharge signal to obtain a time-frequency characteristic atlas, inputs the time-frequency characteristic atlas and sparse dust trajectory mark data into an artificial intelligence discharge type identification model trained by a sparse connection learning framework based on dynamic topology reconstruction and long-interval aggregation to output a discharge starting voltage prediction value, monitors dust fractal dimension by using a dust fractal aggregate parallel computing algorithm to trigger particle size distribution updating, adjusts a dust injection rate according to a deviation between the prediction value and a measured value and records discharge sensitivity reference data, and solves the technical problem of insufficient high-voltage bushing discharge sensitivity evaluation precision under dust invasion conditions.

The present invention proposes a method for estimating the operating state of a power system based on quantum computing, which relates to the field of power system operation analysis and state perception. By constructing a power system operating state estimation model based on an unconstrained quadratic binary optimization model (QUBO model), and solving it through an optical quantum computer, the power system operating state estimation result is obtained. The present invention transforms the power system operating state estimation problem into a QUBO model and solves it, converting the traditional computing algorithm for the state estimation problem into a quantum computing algorithm, and providing a model paradigm and feasible algorithm for constructing quantum computing for the power system operating state estimation problem. This algorithm introduces quantum computing into power system state estimation, and quantum computing will be able to meet higher computing power requirements and shorter response time requirements in the future.

The present invention relates to a partial task offloading and resource allocation method based on LSTM-DDPG, comprising: creating a vehicle network MEC network model for partial task offloading and resource allocation; converting the partial task offloading and resource allocation problems into a reinforcement learning model; and introducing the LSTM neural network into the actor network and critic network of the DDPG algorithm. Compared with the existing DRL-based algorithm, the LSTM-DDPG-based mobile edge computing algorithm of the present invention solves the task offloading and resource allocation problems, can effectively achieve a trade-off between latency and energy consumption, and has significant performance improvements in terms of average task latency and average task energy consumption.

The present invention discloses an online modeling and prediction method for the state information of an unmanned ship based on simulation technology, which relates to the technical field of unmanned ship simulation testing. A database system for the unmanned ship is constructed, including simulation data obtained by simulating an unmanned ship model in a real flow field of star-ccm+ and a Noetic ocean environment, as well as data collected during real sea trials. A nonlinear reservoir computing algorithm based on a sliding window algorithm is designed to perform online modeling and prediction of the state information. By combining the dynamic characteristics of computational fluid dynamics, physical simulation environment, and real ocean environment, the present invention can enable the unmanned ship to realize online modeling and prediction of state information using small sample data in the real ocean environment, provide effective prior information for collision avoidance, planning, and control of the unmanned ship, and safely and effectively complete the navigation task.

A computer system is configured to extend processing executable by a quantum computer in a hybrid system of a classical computer and the quantum computer. The classical computer includes an extraction unit, a generation unit, and a synthesis unit, and the quantum computer includes a control unit and a quantum unit, where the extraction unit extracts a plurality of first element groups as a set of first elements, the generation unit generates parameters based on a first element contained in a corresponding first element group, the control unit controls the quantum unit to cause the quantum unit to execute a quantum computation algorithm on each of the plurality of first element groups based on the parameters and corresponding feature data, and the synthesis unit synthesizes the execution results of the quantum computation algorithm by the quantum unit for each of the plurality of first element groups.

A method (100) for performing event detection on a data stream is disclosed, the data stream comprising data from a plurality of devices connected via a communication network. The method includes using an autoencoder to condense information in the data stream, wherein the autoencoder is configured according to at least one hyperparameter (110) and events are detected from the condensed information (120). The method further includes generating an evaluation of the detected events based on the logical compatibility between the detected events and a knowledge base (130), and using a reinforcement learning (RL) algorithm to improve at least one hyperparameter of the autoencoder, wherein a reward function of the RL algorithm is calculated based on the generated evaluation (140). A system (900) for performing event detection, as well as a method (1100) and a node (1200) for managing an event detection process are also disclosed.

The application provides a time series prediction attack method based on local disturbance. The method comprises the following steps: performing BIM gradient attack on an original sample to obtain a global attack sample; using a sliding window to intercept a local interval of the original sample, performing attack on the interval according to a differential evolution algorithm, converting individuals into time series by using a segmented function, the BIM global attack sample and the original sample to calculate fitness, selecting a better individual to input into a target model to calculate mean square error (MSE) of a prediction value, and continuously iterating the algorithm until the MSE value converges. If the MSE value converges, the local attack sample with successful attack is added to a sequence of attack samples. The application narrows the disturbance range of the attack sample, uses the differential evolution algorithm to find an optimal attack point, and combines the segmented function to segment the disturbance interval, so that the disturbance range is further narrowed, and semi-white-box attack is completed.

A network device controller (104) configured to generate one or more initial Channel State Information, CSI, representative configuration(s) indicating parameters of a CSI representative computation algorithm, transmit the one or more initial CSI representative configuration(s) in one or more first CSI-ReportConfig message(s) to one or more User Equipment(s) (110), UEs, receive a feedback message(s) from at least one of the one or more UEs (110), the feedback message(s) comprising indicators of measurements, perform an evaluation of the measurements and decide if the CSI representative configuration(s) need(s) to be updated, and if so generate an updated CSI representative configuration(s) and transmit the updated CSI representative configuration(s) in one or more second CSI-ReportConfig message(s) to the at least one of the one or more UEs (110).