Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

2788 results about "Batch processing" patented technology

Computerized batch processing is the running of "jobs that can run without end user interaction, or can be scheduled to run as resources permit."

Intelligent short message scheduling method and device based on multi-dimensional dynamic optimization

The invention provides an intelligent short message scheduling method and device based on multi-dimensional dynamic optimization, and the method comprises the steps: obtaining the performance data of a plurality of short message channels, and calculating a channel health score based on a weight dynamic adjustment model; determining a scheduling strategy according to the priority identifier of the to-be-sent message, and performing channel screening and optimal matching; executing message sending and monitoring a sending state; terminal state detection is carried out on the failure message through operator base station signaling, and a decision tree model is applied to determine a retry strategy; and performing Huffman coding compression processing on the P2-level marketing messages which fail in retry, and performing batch sending in an idle window. According to the method, a comprehensive performance evaluation index and reward function model is also constructed, and parameter optimization is performed by applying a reinforcement learning algorithm. According to the invention, multi-dimensional dynamic channel scoring, intelligent retry decision making based on terminal state perception, batch processing with balanced cost-time efficiency and a closed-loop self-optimization system are realized, the short message delivery rate is obviously improved, and the invalid retry rate and the sending cost are reduced.
Owner:BEIJING YULORE INNOVATION TECH

Data processing method and system based on flink unified stream and batch processing, and device and medium

A data processing method and system based on Flink unified stream and batch processing, and a device and a medium, which relate to the field of big data processing. The method comprises: establishing an SQL statement construction rule on the basis of data elements specified in an SQL file standard and a logic relationship between the data elements, so as to perform SQL statement construction on a service to be performed, and obtain a query statement file, wherein the data elements comprise a source table SQL, a destination table SQL and a data conversion SQL; submitting configuration parameters of the SQLs and the query statement file to a Flink cluster, so as to obtain corresponding data processing task information; and the Flink cluster receiving a target data source table on the basis of the data processing task information, processing the data source table on the basis of a predetermined processing mechanism, and then writing corresponding data into a data destination table. The present invention is applicable to different data processing tasks without the data read and write modules in the present invention being rewritten, and it can be ensured that data is executed at least once during a running process of a task.
Owner:CHINA TELECOM BESTPAY CO LTD

Database cascade operation intelligent analysis execution method and system

The invention provides an intelligent analysis and execution method and system for database cascading operation, and relates to the technical field of database management.The intelligent analysis and execution method comprises the steps that a target table and an association table are extracted through a recursive query algorithm, field mapping analysis is carried out to generate a dependency weight matrix, cyclic dependency is recognized based on the matrix and decomposed into directed acyclic subgraphs, and the directed acyclic subgraphs are analyzed; node depth values and breadth values are calculated for topological sorting to generate an execution sequence; performing node classification on the execution sequence, organizing the execution sequence into a batch processing task group, calculating an optimal execution path through an adaptive path optimization algorithm, generating a distributed transaction control instruction, and allocating the task group to an execution thread pool for execution by using a consistent Hash algorithm; monitoring an execution process, dynamically adjusting thread pool resources according to a performance prediction value, executing cascade rollback when an exception occurs, and reconstructing a transaction execution path; according to the method, the execution efficiency and stability of the database cascade operation can be effectively improved, and the system resource consumption is reduced.
Owner:NORTH CHINA MUNICIPAL ENG DESIGN & RES INST

Scaling utilization of large language models

The present disclosure relates to efficiently receiving and processing input tasks in a way that is scalable and which reduces both the quantity of tokens processed by a foundation model (e.g., an LLM) as well as the number of API calls that are made in processing the input tasks. A system batches a set of inputs to provide as a single batch of input(s) into an LLM. The system generates one or more permutations of the batched input(s) to determine outputs based on variable orders in which the input data is provided within the respective permutations of the batched inputs. The system further may eliminate one or more of the data inputs within the respective batches to facilitate smaller batched inputs without sacrificing accuracy in a set of outputs generated by the LLM responsive to the batch permutations.
Owner:MICROSOFT TECHNOLOGY LICENSING LLC

Adaptive Data System And A Method For Cognitive Data Processing

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.
Owner:DATAQUANTUM INC

Communication transmission control cabinet based on big data

The invention discloses a communication transmission control cabinet based on big data, and relates to the technical field of communication transmission, and the control cabinet comprises the steps: introducing a real-time network state perception and prediction module, collecting the data of a link through an embedded sensor, carrying out the prediction through combining with an LSTM time sequence model, carrying out the estimation of a future link state, introducing a multipath transmission mechanism, and carrying out the real-time network state perception and prediction. When tasks are distributed, bandwidth, delay and load conditions of paths are synthesized, batch uploading is performed through caching and compression strategies of edge computing nodes, occupation of links is effectively reduced, when link performance is reduced, a path switching mechanism can select alternative paths with optimal performance in real time, influences of link interruption on communication tasks are avoided, and meanwhile, the communication efficiency is improved. According to the high-real-time task priority strategy, an optimal path is adopted, delay of key tasks is reduced, non-real-time tasks are compressed and processed in batches at edge nodes, resources are not competed with real-time tasks, interference is avoided, and the requirements for ship navigation and real-time monitoring tasks are met.
Owner:HEBEI SHOUKE YUANTUO TECH CO LTD

Three-dimensional scene loading and rendering method for deep sea three-dimensional environment deduction

The invention discloses a three-dimensional scene loading and rendering method for deep-sea three-dimensional environment deduction, and relates to the technical field of three-dimensional scene rendering, and the method comprises the following steps: obtaining and preprocessing various types of data of a deep-sea environment, and carrying out the grid processing of terrains and objects in the deep-sea environment; a deep sea three-dimensional scene is established based on the preprocessed deep sea environment data, scene uniform segmentation is carried out, a plurality of sub-scene blocks are obtained, and a block rendering sequence table is obtained; the rendering efficiency of the deep sea three-dimensional scene is remarkably improved by adopting the view cone removing technology and the batch processing rendering strategy, the rendering speed is increased, dependence on hardware resources is reduced, the rendering task can be efficiently completed under the condition that the resources are limited, and in addition, the rendering efficiency of the deep sea three-dimensional scene is improved. The strategy of dynamically adjusting the rendering detail level further balances details and performance, and ensures that the optimal rendering effect can be provided under different observation distances and hardware loads.
Owner:NAT DEEP SEA CENT

Label-based user portrait adaptive updating method, medium and equipment

The invention discloses a label-based user portrait self-adaptive updating method, a medium and equipment, and solves the problems of poor data timeliness and lagging rule updating in a traditional offline batch processing mode through a collaborative mechanism of real-time stream processing and historical batch processing. A dynamic weight adjustment strategy is adopted, real-time user tags generated by real-time behavior flow data and batch user tags generated by historical behavior batch data analysis are subjected to aging weighted fusion, and dynamic changes of user behavior modes are effectively captured; through feedback behavior data of the user and a closed-loop iteration mechanism of the rule base, self-adaptive evolution of the label rule is realized, and the accuracy and the response speed of a recommendation system are remarkably improved.
Owner:ZHONG FU TONG CO LTD

Multi-NDP system read-write acceleration method based on key value separation technology

The invention discloses a multi-NDP (Near-Data Processing) system read-write acceleration method based on a key value separation technology, and the method comprises the steps: when data is written, splitting a data key value pair written into a batch processing region of a multi-NDP system, determining target equipment, then carrying out key processing and value processing, carrying out key storage on keys, and carrying out data storage on the target equipment; a value corresponding to the key is added to a value log file of the target equipment; during data query, hash operation is carried out on a query key, target device position information of a corresponding value is obtained, then the master-slave multi-level hot cache is queried, and query retrieval of the corresponding value is carried out according to the obtained value position index. According to the method, the key value separation technology is introduced into the multi-NDP system, when the Compaction is executed, the execution time of the Compaction is shortened, the influence of the Compaction on the writing operation of the user is reduced, and the writing performance is greatly improved.
Owner:ANHUI UNIV

System and method for hybrid processing of mass offline data and mass real-time data

The invention relates to the technical field of electric data processing, and discloses a system and a method for hybrid processing of mass offline data and mass real-time data, and the system comprises a queue management module which receives task requests, identifies and marks task types, distributes tasks to a stream processing queue or a batch processing queue, and sorts the tasks according to priorities; the task types comprise flow tasks and batch tasks; the thread pool management module establishes a thread pool, sequentially processes tasks according to priorities and collects real-time parameters; the calculation module is responsible for calculating task priorities and calculating task jitter factors based on real-time parameters; the scheduling decision module compares the task jitter factor with a reference value, determines a weight adjustment coefficient, corrects the priority and formulates a scheduling strategy; the alarm module judges whether to trigger an alarm. According to the invention, by introducing the dynamic priority scheduling and preemptive processing mechanism driven by the jitter factor, the real-time response capability and the resource utilization efficiency of the system are improved.
Owner:BEIJING LIUJINSUIYUE TECH CO LTD

Large language model reasoning service optimization method, system, equipment and medium

The invention discloses a big language model reasoning service optimization method, system and device and a medium. The method comprises the steps of obtaining a current reasoning request queue; performing dynamic sorting on each reasoning request in the current reasoning request queue; and performing dynamic batch processing on each reasoning request in the reasoning request queue by using an SLO-BS algorithm to obtain a current batch of reasoning requests, and inputting the current batch of reasoning requests into a large language model, the method, the system, the equipment and the medium can solve the problem of video memory fragmentation occurring during multi-task concurrent reasoning, and reduce the SLO violation rate at the same time.
Owner:SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI

Dynamic batching of prompt queries

A model comprises different processing nodes that perform different functions on input data, such as data prompts. Each node has batching criteria including a threshold value that when exceeded triggers a transmission of a batch of processing requests to the corresponding node. Different processing requests for a node are routed to a batching cache for that node. Processing requests are prevented from being routed to the node until the batching criteria is met. Once the batching criteria is met, a batch of processing requests stored in the batching cache are routed to the corresponding node.
Owner:MICROSOFT TECHNOLOGY LICENSING LLC

Rural water supply pipe network pressure monitoring analysis method and device based on digital twinning

The invention discloses a rural water supply pipe network pressure monitoring analysis method and device based on digital twinning. The method comprises the following steps: collecting historical basic data of a rural prediction area; flow and pressure sensors are arranged at nodes of a rural water supply network to obtain monitoring data; uploading the data to a server through a data stream processing framework, and preprocessing the data to obtain preprocessed data; constructing a relational database management system to perform batch processing on the preprocessed data to obtain historical basic data and monitoring data after batch processing; constructing a rural pipe network water pressure prediction model; constructing a pipe explosion detection and water pressure control model; according to the physical data, constructing a rural water supply pipe network geometric model based on ArcGIS; and constructing a rural water supply pipe network system data model and realizing visualization of the operation state of the water supply pipe network. Rural water supply characteristics and water consumption characteristics are fully considered, intelligent management and control of the rural water supply system are achieved, and reliability and safety of rural water supply are effectively guaranteed.
Owner:HUNAN PROVINCIAL WATER CONSERVANCY & HYDROPOWER SURVEY & DESIGN INST GENERAL INST

Machine learning model scalability with distributed multi-layer processing

Machine learning model scalability with distributed multi-layer processing is disclosed herein. A method for processing and deploying machine learning models that enhances scalability and efficiency by executing a subset of a neural network on each of a plurality of interconnected processing units. The method involves partitioning compute tasks across these processing units to reduce latency, including broadcast and reduction processes for inputs and outputs. It also includes managing the allocation of samples in a batch to specific master processing units within the distributed arrangement and synchronizing computation between fully connected layers within each processing unit. Additionally, the method implements data reduction during the transfer of data across the processing units, wherein data is accumulated with a current processing unit's partial sum as it is transferred to the destination processing unit.
Owner:EXPEDERA INC

Network security information analysis and statistics method based on regulation and control cloud platform

The invention discloses a network security information analysis and statistics method based on a regulation and control cloud platform. Protocol identification, standardized mapping and blood relationship marking of multi-source security equipment data are realized through an edge layer protocol adaptation cluster and a metadata self-registration mechanism; constructing a real-time streaming and batch processing dual-channel framework based on regulation and control of cloud platform container resources, and fusing to generate a standardized security event object with a frequency weight and a hazard coefficient; designing a main ring-emergency ring double-path decision closed loop, respectively processing conventional events and high-risk threats, and realizing knowledge base iteration; extracting features by using a containerized architecture in combination with an XGBoost + isolated forest integrated model, detecting threats and generating rules; and a space-time-system-risk four-dimensional visualization engine is constructed, attack path drawing, event evolution display and high-risk area prediction are realized, and a management closed loop is formed. According to the network security information analysis and statistics method based on the regulation and control cloud platform, the intelligence and automation level of network security information analysis and statistics is improved.
Owner:INFORMATION & COMM COMPANY OF QINGHAI ELECTRIC POWER +1

Cross-border e-commerce big data analysis system and method

The invention provides a cross-border e-commerce big data analysis system and method, and the method comprises the steps: obtaining multi-dimensional original business data of a cross-border e-commerce platform, carrying out the real-time collection through a distributed data collection system, and obtaining a structured business data set; data collected in real time is cleaned, converted and structured, a unified data model is constructed, and integration of heterogeneous data sources is achieved; storing the processed structured data in a time sequence database to provide data support for real-time analysis; performing real-time aggregation and statistical analysis on the newly added data to generate business index and key performance index data; a Lambda framework is adopted, batch processing and stream processing are combined, the historical data and the real-time data are subjected to full-amount analysis, and an analysis result is obtained; based on an analysis result, a data mining algorithm is adopted to identify a service mode and predict a future trend; a multi-dimensional data cube is constructed, multi-dimensional and multi-granularity data exploration and interactive analysis are supported, and management decision making is assisted.
Owner:CHENGDU TAILYOU TECHNOLOGY CO LTD

Big data secure storage method and system based on cloud computing

The invention relates to the technical field of data security, in particular to a big data security storage method and system based on cloud computing, and the method comprises the following steps: initializing a cloud service node, selecting data blocks needing security storage, distributing an identifier for each data block, sorting and classifying the data blocks by adopting a batch processing mode, and storing the sorted data blocks into a database; and obtaining an initialized data block. According to the method, the security of the data in the cloud environment is greatly improved through key segmentation and storage of part of keys in a plurality of security cloud service nodes, the decryption rule is constructed by using the threshold cryptography model, the data can be decrypted only when a sufficient number of key parts are collected, unauthorized access and data leakage are effectively prevented, and the security of the data in the cloud environment is improved. The strategy of dynamically adjusting the secret key replacement period adapts to the access frequency and the change dynamic state of the storage data block, and the confidentiality, the integrity and the availability of the data are improved, especially in a sensitive service scene with a large data volume.
Owner:BEIJING MUXUE COMPUTER TECHNOLOGY CO LTD

CPU (Central Processing Unit), GPU (Graphic Processing Unit) and NPU (Network Processing Unit) resource allocation method and system for training and calculating integrated machine

The invention relates to the field of heterogeneous computing resource management, in particular to a CPU (Central Processing Unit), GPU (Graphic Processing Unit) and NPU (Network Processing Unit) resource allocation method and system of a training and reckoning all-in-one machine. The invention discloses a CPU, GPU and NPU resource allocation system of a training and reckoning all-in-one machine. The system comprises a schedulable resource analysis module, a load analysis module and a resource allocation module. According to the method, the reference parameters and the real-time dynamic indexes of heterogeneous hardware are fused, so that deep perception and efficient quantification of computing power resources are realized; on the basis of an NPU temperature attenuation experiment, calibrating a computing power loss coefficient, and dynamically correcting available weight values of a CPU, a GPU and an NPU in combination with suitability analysis of a GPU stream processor utilization rate and a task batch processing demand and collaborative efficiency evaluation of a CPU dominant frequency and an IPC value; and the accuracy of resource availability prediction is improved, so that the system can still guarantee the stability of computing power output in complex environments such as high temperature and high concurrency.
Owner:DONGGUAN HUAMING TENG TECH CO LTD

Intelligent laboratory full-process collaborative management system and method based on multi-dimensional data fusion

The invention provides an intelligent laboratory full-process collaborative management system and method based on multi-dimensional data fusion, and the method comprises the steps: obtaining an original data stream from laboratory instrument equipment, carrying out the format recognition and conversion of heterogeneous instrument data through a standard protocol adapter, and generating a first data set in a unified format; aiming at the first data set, executing automatic analysis by adopting data middleware, and integrating into a structured second data set based on a metadata rule of field mapping and unit conversion; according to the second data set, a multi-source data fusion model is constructed, data are classified according to instrument types through batch processing and aggregated according to timestamps, and a fused third data set is generated; aiming at the initial process configuration, adjusting a node sequence and a parameter threshold value by using a visual design tool, and generating an optimized process configuration adaptive to the diversified scene; according to the optimized process configuration, process scheduling is executed through a dynamic process engine, execution time and resource occupation are monitored in real time, and a process execution log is generated.
Owner:HUNAN WEIBO INFORMATION TECHNOLOGY CO LTD

Multi-head attention mechanism operator optimization method and system for domestic heterogeneous processor

The invention discloses an optimization method for a multi-head attention mechanism operator oriented to a domestic heterogeneous processor, which improves the efficiency of the multi-head attention mechanism operator from four aspects, and comprises the following steps of: firstly, combining the general computing power of a CPU (Central Processing Unit) and the advantages of a DSP (Digital Signal Processor) in the aspect of parallel computing; a high-performance multi-head attention mechanism operator operation method on a CPU + DSP heterogeneous processor is constructed; secondly, a previous softmax optimization algorithm is used for fusing the two batch processing matrix multiplication and a softmax operator, so that the memory access overhead of DSP equipment is reduced; then, a vector operation instruction of a DSP calculation core is used in the calculation process of a multi-head attention operator, and SIMD calculation characteristics of DSP equipment are fully utilized; and finally, during data transmission and calculation, a double-flow water mechanism is adopted, the data transmission time is covered, and the multi-head attention mechanism operator calculation efficiency is improved.
Owner:HUNAN UNIV

Scheduling optimization method for aviation composite material manufacturing hot press molding process

The invention discloses a scheduling optimization method for an aviation composite material manufacturing hot-press forming process, which comprises the following steps of: 1, establishing a hot-press forming production plan scheduling objective function by taking the minimum total delay time of a workpiece as an objective according to an aviation composite material manufacturing hot-press forming production plan; 2, establishing a two-dimensional hot press molding process batch scheduling model considering a workpiece release time window and machine applicability; and step 3, taking the objective function as an optimization objective, taking the two-dimensional hot-press forming process batch scheduling model as an optimization model, and utilizing a genetic algorithm embedded with a self-adaptive domain search strategy to solve and obtain a hot-press forming production plan scheduling scheme. According to the method, the total delay time of workpieces can be remarkably shortened while the machining batch of the workpieces is minimized, the occurrence of prepreg deterioration conditions is reduced, and conditions are created for improving the hot press molding production efficiency and promoting digital and intelligent transformation and upgrading of composite material manufacturing enterprises.
Owner:NANJING UNIV OF AERONAUTICS & ASTRONAUTICS

Parameterized machine learning pipeline implemented using a lambda architecture

A method for developing and deploying a machine learning pipeline is provided. The method may be implemented by at least one processor. The method may include collecting different types of data from one or more data sources, wherein the different types of data includes batch data and real-time data; continuously processing the data to build a plurality of layers of a Lambda application, wherein each layer is configured to be deployed independently of the other layers; building a modular machine learning pipeline in a modifiable reference file, wherein the modular machine learning pipeline defines a framework comprising a no-code / low-code parameterized machine learning framework; deploying the modular machine learning pipeline as a single deployment machine learning pipeline that can be modified for different use cases by revising the source code of a CI / CD (Continuous Integration and Continuous Delivery / Continuous Deployment) pipeline, while the framework of the machine learning pipeline remains unchanged.
Owner:JPMORGAN CHASE BANK NA

Reducing latency through propensity models that predict data calls

There are provided systems and methods for reducing latency through propensity models that predict data calls. A service provider, such as an electronic transaction processor for digital transactions, may provide computing services to users including those for electronic transaction processing. In order to provide sequence-based forecasting of computing events and processing requests for users, accounts, and / or activities associated with the service provider, the service provider may provide a machine learning model, such as a deep neural network, that predicted potential occurrences and likelihoods of computing events occurring at future times. When predicting the events, the service provider's machine learning predictive framework may further predict data calls required to be executed to retrieve data needed for processing during the events. These predicted calls may then be batched together into a batch processing job, which may be executed to retrieve the data prior to the predicted events.
Owner:PAYPAL INC

Excel report automatic conversion and rendering method, system, equipment and medium

The invention discloses an Excel report automatic conversion and rendering method, system, device and medium, and relates to the field of data processing, visualization and document generation, the method comprises the following steps: adopting an intelligent Excel analysis engine to automatically analyze Excel file structure features, and selecting a most suitable analyzer to analyze an Excel file; based on an xlsx library and a data structured conversion mechanism, performing data processing on the analyzed Excel file, and constructing a structured data model; based on a dynamic component of React, constructing different types of visual generation components according to the structured data model, and determining a layout framework; and based on a dynamic component and a structured data model of React, dynamically loading and rendering the visual content in the layout framework by adopting a headless browser technology, configuring browser parameters suitable for generating a PDF format, and displaying a PDF document according to the browser parameters. The method has the advantages of high-flexibility Excel analysis capability, strong data processing capability, rich visual expressive force, high-quality PDF rendering and efficient batch processing capability.
Owner:BEIJING QINGWANG TECH CORP

Task processing method and device, computer equipment and storage medium

The invention relates to the technical field of artificial intelligence and natural language processing, and discloses a task processing method and device, computer equipment and a storage medium, and the method comprises the steps: carrying out the dynamic batch processing of an input text task, and obtaining a conventional batch and an ultra-long batch; detecting a task belonging to an ultra-long batch in the to-be-processed text task as a priority text task, and realizing preemptive scheduling of the current text task by the priority text task through a hardware interrupt mechanism; calculating the priority text task by adopting a local-global mixed attention mechanism; calculating the text tasks belonging to the conventional batch by adopting a first sparse rate, and generating a first historical key value pair; calculating a priority text task by adopting a second sparse rate, and generating a second historical key value pair; and caching the first historical key value pair in a first cache pool, and caching the second historical key value pair in a second cache pool. The method can be applied to the field of financial science and technology businesses, and efficient differentiation processing and computing resource optimization of super-long and conventional text tasks are achieved.
Owner:PING AN TECH (SHENZHEN) CO LTD

Elastic converter service system self-adaptive through lexical elements

The present disclosure relates to methods and systems based on a transformer model, in which an elastic transformer service system, referred to herein as an online lexical adaptation system (OTAS), is involved that accommodates diversified user requests with fluctuating query loads, while optimizing output accuracy and runtime delay. The OTAS uses lexical adaptation techniques that involve adding cue lexical elements to improve accuracy and removing redundant lexical elements to accelerate reasoning. To handle fluctuating query loads and diversified user requests, the OTAS also applies a perceptual selective batch processing algorithm for online lexical adaptive design. In a system service process, OTAS firstly processes incoming queries with similar requirements in batches to improve ingress throughput. Then, in order to balance between the overhead of lexical element growth and the potential of improving accuracy, OTAS adaptively adjusts lexical element execution settings by solving an optimization problem.
Owner:THE HONG KONG UNIV OF SCI & TECH

Substrate processing system, substrate processing method, and recording medium

A substrate processing system includes a carry-in / out unit in which a cassette accommodating therein multiple substrates is carried in and out; a batch processing unit configured to process a lot including the multiple substrates at once; a single-wafer processing unit configured to process the substrates one by one; a first interface unit configured to distribute the substrates to the single-wafer processing unit or the batch processing unit; and a second interface unit configured to transfer the substrates between the batch processing unit and the single-wafer processing unit. The first interface unit includes a first placement unit configured to place therein the substrates before and after being processed by the single-wafer processing unit; a second placement unit configured to place therein the substrates before being processed by the batch processing unit; and a transfer device configured to transfer the substrates to the first placement unit and the second placement unit.
Owner:TOKYO ELECTRON LTD

Edge node scheduling system, edge node scheduling method and edge node

The invention relates to the field of wireless communication, in particular to an edge node scheduling system, an edge node scheduling method and an edge node.The edge node scheduling system comprises a data request module, a management back-end module, a performance collector, a scheduler, a predictor and a controller, classifying and putting into a dynamic queue; the management back-end module calls a concurrent reasoning model to process data in batches according to resource and task requirements, and generates a network performance index; the performance collector continuously monitors and collects the indexes and the resource utilization rate; the scheduler formulates a node resource scheduling strategy accordingly; the predictor predicts resource interference and assists the scheduler to make a decision; and the controller switches the model according to the scheduling strategy, adjusts batch size and concurrency number, and realizes dynamic optimization of resources. Through the innovative edge node scheduling framework, localized efficient processing of the data is realized, the execution efficiency of the AI reasoning task is improved, and the resource utilization rate is remarkably improved.
Owner:JIANGNAN UNIV +1

Distributed streaming multi-mode fusion adaptive gradient compression optimization method and system

The invention belongs to the field of multi-modal RGB-T data fusion, and discloses a distributed streaming multi-modal fusion-oriented adaptive gradient compression optimization method and system, and the method comprises the steps: carrying out the personalized compression of data features of different modals through employing a modal-sensitive gradient compression strategy, and enabling the modals to comprise an RGB modal and an infrared modal; introducing physical constraint time sequence alignment loss, and performing time sequence alignment on the multi-modal data subjected to personalized compression by utilizing a physical model based on a thermal diffusion equation; the dynamic batch processing strategy is sensed through the video memory, the use condition of the video memory is monitored in real time, and the batch processing size is dynamically adjusted according to the residual capacity of the video memory. According to the method, a brand new solution is provided for efficient training of distributed streaming multi-modal data, and the precision of a multi-modal data fusion task and the resource utilization rate are improved.
Owner:CHONGQING NORMAL UNIVERSITY

YOLOv8-based marine ship target rapid identification method

The invention discloses a rapid marine ship target identification method based on YOLOv8, and is applied to the technical field of target identification. Comprising the following steps of collecting an image sample of a ship navigation water area in a digital vehicle, cleaning and marking the image sample, performing data enhancement on the image sample to obtain a data set which can be used for training and verification, constructing a target identification network model based on a YOLOv8 deep neural network, and obtaining a target identification result by setting hyper-parameters such as a learning rate and a batch processing size. And carrying out transfer learning by using the pre-training model, removing a repeated detection frame by using a non-maximum suppression algorithm and carrying out result filtering by using a confidence coefficient threshold to carry out post-processing on a reasoning result of the model so as to complete prediction of the ship target, and carrying out sparse pruning training on the trained model neural network so as to obtain the ship target. And carrying out calibration and quantification on the pruned model. According to the method, on the premise that the model detection precision is reserved to the maximum, memory consumption during equipment operation is reduced, and meanwhile the model detection speed is increased.
Owner:ZHEJIANG UNIV OF SCI & TECH