245 results about "Storage model" patented technology

A resource set which includes multiple servers with a respective plurality of training computing devices is identified for training a machine learning model. The resource set is subdivided into partition groups, such that each partition group can store a respective replica of state information of the model. The model is trained using the partition groups. The training comprises a multi-stage gathering of a portion of the state information at training computing devices of a particular partition group. Different types of communication channels between training computing devices are used in respective stages of the gathering, including inter-server communication channels in one stage and an intra-server communication channel during another stage. A trained version of the model is stored.

The present invention discloses a method for constructing an intelligent computation engine of an artificial intelligence cross-platform model based on knowledge self-evolution. The method comprises: determining source and target moments; dividing a discrete manufacturing system data set; initializing a dynamic discrete manufacturing system model; preprocessing data and constructing a task pool; constructing a meta learning framework; migrating a trained neural network to new tasks; iterating until convergence and storing model parameters; and testing in a new environment. This invention shortens the convergence time of model parameters, significantly benefiting the training of dynamic discrete manufacturing models subject to temporal disturbances in actual production.

Methods and apparatus for estimating an energy storage performance property of an energy system are provided. The method includes providing, as input to an energy storage model and a machine learning model, sensor information associated with the energy system, wherein the energy storage model comprises an empirical model and / or a physics-based model, providing as input to the machine learning model, one or more values based on an output of the energy storage model, and determining based, at least in part, on an output of the machine learning model, an estimate of the energy storage performance property.

An apparatus of a user equipment (UE), the apparatus comprising a processor, and a memory storing instructions that, when executed by the processor, configure the apparatus to receive, from a base station of an operator network, Protocol Data Units (PDUs) carrying Artificial Intelligence (AI) model data in either a control plane or a user plane, and decapsulate the PDUs to obtain and store the AI model data, wherein the AI model data is indicative of an AI model configured for inference in Access Network (AN) protocol layers at the UE.

The invention relates to the technical field of automatic warehouse material, and discloses an intelligent goods allocation distribution system for automatic warehouse logistics, and the system comprises a goods allocation data collection module; a dynamic state detector; a multi-source data fusion unit; a three-dimensional storage modeling module; intelligently distributing a strategy engine; and a PLC control interface. According to the intelligent goods allocation allocation and scheduling system for automatic warehouse logistics, a goods allocation allocation scheme is generated through a multi-target Pareto optimization algorithm, the balance between carrying distance minimization min sigma alpha * Di and space utilization rate maximization Ut is achieved, and compared with a proportional static allocation rule, the warehouse space utilization rate is increased to 82%; a dynamic grid map technology is integrated, and a track cross risk area (red early warning grid) of the stacker and the AGV is marked in real time, so that the equipment collision probability is reduced by 90%; and in combination with a goods relevance knowledge graph, special goods such as medicines are clustered and stored according to temperature zones, and the energy consumption of cross-temperature-zone operation of equipment is reduced by 35%.

The embodiment of the invention provides a code compiling method, electronic equipment and a storage medium. The code compiling method is used for compiling a code comprising at least one thread local storage variable, and comprises the following steps: in response to existence of a target thread local storage variable, allocating the at least one target thread local storage variable to a thread local memory based on an attribute of the at least one target thread local storage variable, obtaining a thread local memory address corresponding to each target thread local storage variable, wherein the target thread local storage variable is the thread local storage variable used by the first function; updating an intermediate representation corresponding to the code based on a thread local memory address corresponding to each target thread local storage variable; and generating a machine code corresponding to the code based on the updated intermediate representation. According to the code compiling method, thread local storage model support under a parallel computing architecture is provided, and the problem that traditional thread local storage implementation cannot be directly applied to a GPU architecture is solved.

Automated machine learning (Auto ML) for creating and optimizing ML models using a model store for storing: trained ML models and hardware models; test metrics data corresponding to the stored models; ML advised-models. Using a model meta-services (MMS) for: accessing the stored models and the test metrics data; creating the ML meta-models based on the runtime test metrics data; and answering MPC queries. Using a models producer and consumer (MPC) for: selecting a ML advised-model; testing the selected ML advised-model using selected ML test inputs and outputs to provide runtime test metrics data; optimizing the selected ML advised-model using the runtime test metrics data; sending the optimized ML advised-model to the model store unit for storing as one of the stored ML advised-models; and sending the runtime test metrics data to the model store unit for storing as part of the runtime test metrics data; and sending the MPC queries.

The invention discloses a multi-dimensional data asynchronous calculation method and system based on a dynamic dependency graph, and relates to the field of data processing. The method comprises the following steps: constructing a metadata, logic and instance three-layer separation storage model; analyzing the reference relationship to construct a directed acyclic graph, and generating a calculation priority; monitoring data change, and executing asynchronous serialization calculation through a message queue based on priority; the associated document is automatically updated based on anchor mapping. According to the method, the calculation deadlock and the performance bottleneck of large-scale data in the Web environment are solved, logic decoupling and dynamic expansion are realized, the final consistency of the data is guaranteed, and the high-concurrency throughput and the stability are remarkably improved.

The invention discloses a supply chain logistics real-time scheduling system and method, and relates to the technical field of production scheduling, and the system specifically comprises a production plan dynamic adjustment module, a digital twin simulation module, a dynamic response and optimization module, a space-time coupling resource scheduling model, a cross-enterprise cooperation module, and a visualization and decision support module. PLTI instructions are dynamically generated through digital twinborn simulation and a dynamic response mechanism, in-plant and out-plant real-time collaboration and in-plant and out-plant combined modeling are achieved, the inventory turnover rate is increased by 25%, the inventory turnover rate is increased through a space-time coupling resource scheduling model, an in-plant storage model and an Amap API road network are constructed through a Unity 3D engine, instruction conflict probability is rehearsed, and the real-time performance of the system is improved. According to the method, logistics scheduling information is dynamically updated every 15 minutes based on a production plan, the logistics scheduling conflict probability is reduced by 83% through digital twin rehearsal, and real-time collaboration and intelligent scheduling of logistics inside and outside a factory are achieved through digital twin simulation, a dynamic response mechanism and a space-time coupling resource scheduling model.

The invention discloses an electroencephalogram voice annotation calibration and scheduling system and method for individual differences. The system comprises an electroencephalogram signal collecting and preprocessing module, a voice instruction receiving and recognizing module, a rapid individual calibration module, a dynamic task scheduling module, an electroencephalogram feature online decoding and annotation module and an annotation storage and feedback module. Through cooperative work of the two core modules, namely the rapid individual calibration module and the dynamic task scheduling module, the problems that decoding performance is reduced during cross-subject and cross-task switching caused by individual differences, and multi-task labeling efficiency is low due to task scheduling strategy stiffness are solved accurately. Data persistent storage, model online updating and user interaction feedback are completed through storage, optimization and feedback links, periodic retraining is carried out by utilizing daily successful labeling data through a model continuous learning link, and finally, a zero-threshold individual rapid adaptation and high-efficiency multi-task collaborative labeling target is achieved.

The invention discloses a multi-source heterogeneous big data aggregation analysis method based on master data modeling. The method comprises the following steps: modeling main data, defining a relation mode of a business entity and a data entity, and outputting a graph structure storage mode; classifying the service entities, and defining different data entity fusion algorithms for different data entity types by taking the main attributes of the service entities as the identification basis of the data entities; accessing data, and cleaning the accessed data; service entities and data entities are analyzed from the cleaned data based on a data entity fusion algorithm, and a relation theme library is constructed according to the incidence relation between the service entities and the data entities; storing the cleaned data and a relation theme library thereof; and carrying out modeling analysis on the cleaned data by combining the characteristics of the business entity and the relation theme library. According to the method, unified management of multi-source heterogeneous mass data is carried out from a global perspective, the blood relationship of the data is clear, and the authority is controllable.

The invention provides an Internet of Vehicles data sharing and storage model based on a double-layer block chain, and aims to solve the problems of low data interaction efficiency, poor expansibility, insufficient security and the like in the traditional Internet of Vehicles. According to the model, road side units are grouped geographically, an optimized Raft protocol is adopted in each group to realize rapid consensus, and an improved PBFT protocol is adopted among the groups to ensure global consistency. Self-adaptive leader election, batch processing and assembly line mechanisms are introduced into the Raft protocol, so that the throughput and response speed of the system are improved; a reputation value-based VRF main node election mechanism and a BLS aggregation signature technology are introduced into the PBFT protocol, so that the communication complexity is reduced, and the system security and fault-tolerant capability are enhanced. According to the method, the fault-tolerant rate and the message complexity of the system are analyzed theoretically, the advantages of the system in the aspects of throughput, time delay and fault-tolerant performance are verified through simulation experiments, and the method is suitable for large-scale and high-concurrency car networking application scenes.

The invention discloses a reasoning acceleration method and device of a pre-training model and electronic equipment, and relates to the technical field of computer application, the method comprises the steps of preprocessing a behavior sequence, combining the preprocessed behavior sequence with a neighbor sequence, inputting the preprocessed behavior sequence into the pre-training model, effectively extracting a dependency relationship through the design of a self-attention layer and a graph convolution layer, and improving the reasoning speed of the pre-training model. Performing aggregation processing to realize accurate reasoning of the pre-training model; meanwhile, the intra-core register which is located in the calculation core and used for storing the model parameters is arranged in the programmable logic unit, and the model parameters which are frequently used in the pre-training model are stored in the intra-core register, so that the reasoning efficiency of the pre-training model is improved.

The invention provides a micro-grid black start recovery optimization method and device based on adaptive Benders decomposition, and the method comprises the steps: providing a Benders two-stage solving algorithm fused with an adaptive penalty factor for the hierarchy and constraint complexity of a micro-grid black start two-stage optimization model. In the first optimization stage, a micro-source recovery sequence and a recovery path are optimized based on a Benders decomposition method of an adaptive penalty factor, and in the second optimization stage, load recovery is optimized in combination with mixed integer programming. According to the method, the precision of the optical storage model is improved through dynamic characteristic correction, the recovery efficiency and stability are balanced through a two-stage framework, the solving performance is improved through an adaptive algorithm, and efficient and safe recovery of black start of the micro-grid is effectively achieved.

The invention belongs to the technical field of data processing, and particularly discloses a computer vision model service scenarized configuration generation method, which comprises the following steps: constructing a basic information database comprising a plurality of computer vision models, the database being used for storing model identification information and feature parameters; extracting structured features of the scene according to a scene analysis engine, wherein the structured features are used for describing operation conditions of the scene; calculating a model matching score according to the structured features and the feature parameters, and obtaining a recommendation model through a multi-condition weight matching function; executing a model selection strategy based on the matching score, and obtaining a model matched with the scene; and generating a configuration file conforming to a predetermined specification according to the recommendation model. The objective of the invention is to solve the problems of low efficiency and error proneness caused by the fact that a computer vision model in the prior art depends on a manual selection model and cannot quickly select and adapt to an optimal model in diversified scenes.

The invention discloses a multi-dimensional index calculation method based on a graph database. The method comprises the following steps: defining an index graph storage model based on the graph database; registering a parameterized calculation task corresponding to the index in a built-in calculation framework of a graph database; when a user query request arrives, submitting a query transaction through a long connection channel established by driving the graph database, and directly accessing local storage data by a calculation framework to complete real-time calculation and return a result; when the original data in the graph database is changed, a database trigger captures a change event and pulls up the calculation task in the same process; user query triggering and data updating triggering share the same set of calculation task and graph traversal logic, and storage and calculation integrated real-time multi-dimensional index calculation and updating are achieved. The multi-dimensional index calculation method based on the graph database provided by the invention only depends on the graph database and its own calculation framework, simplifies a data processing link, and improves the index output efficiency.

The invention discloses an intelligent patrol control method, and relates to the technical field of intelligent patrol, and the intelligent patrol control method mainly comprises the steps: obtaining a resource data storage model according to patrol resources; according to the resource data storage model, a quick response police force task is obtained based on shortest path calculation and a task matching method; a major alarm condition or a safety event is confirmed, and emergency response processing is performed according to the quick response police force task; and performing multi-device cooperative scheduling according to the scheduling task and the patrol resources. By implementing the intelligent patrol management and control method provided by the invention, the police scheduling efficiency and response capability can be improved, the accuracy and collaboration of police resource integration can be improved, the system construction and operation cost can be reduced, the safety and reliability of the system can be enhanced, and the handling capability of emergency events can be improved.

The invention discloses a special tariff business settlement method and device, equipment, a storage medium and a product, and relates to the technical field of information technology data support, and the special tariff business settlement method comprises the steps: when a settlement instruction is received, extracting corresponding special tariff information from a real-time data warehouse, the real-time data warehouse synchronizes the service data of the special settlement service database in real time, the service data is obtained by performing data extraction on the obtained ticket file, and the service data is fragmented and stored to the special settlement service database through a distributed storage model loading fragmentation rules from a data source; and settling the special tariff information to obtain a special tariff settlement report. According to the invention, the distributed storage model is adopted to store the service data in the special settlement service database in a fragmented manner, so that super-large-scale file data volume can be borne; the special tariff information is extracted from the real-time data warehouse, the special tariff settlement report is output through automatic settlement, and the settlement efficiency of the special tariff is improved.

The invention discloses an efficient storage method and system for distributed molecular dynamics simulation calculation, and the method comprises the steps: constructing a distributed storage model based on data access domain division, and dividing the data stored in a node into private access data and shared access data; a storage structure with particle multi-dimensional information separation is adopted, particle data is divided into static data and dynamic data, and different storage structures are adopted for different types of data. A hierarchical data mapping and addressing method is adopted, storage and access of particle data are divided into a simulation space level, a calculation node level and a space grid level, each layer is organized through a specific numbering rule, and effective distribution and efficient access of the particle data are ensured. According to the efficient storage method and system for distributed molecular dynamics simulation calculation, the problems of uneven particle data distribution and low access efficiency in the molecular dynamics simulation process are solved.

The invention discloses a standard metadata storage method and a machine readable storage medium, and the method comprises the steps: collecting operation data and associated data of a preset coal chemical industry, and carrying out the preprocessing of the equipment data and the associated data; performing multi-modal semantic information extraction on the associated data to obtain information data, performing structured processing on the information data and the operation data to obtain structural data, and constructing a process mapping table according to business terms and the process parameters; the information data comprises incremental information and general information; constructing a coal chemical engineering topological structure knowledge graph according to the structure data and the process mapping table based on a chemical engineering association type, and performing time sequence coding on the coal chemical engineering topological structure knowledge graph according to a time window and a process unit to obtain a dynamic code; and constructing a coal chemical industry standard metadata storage model according to the dynamic codes, inputting to-be-stored data into the coal chemical industry standard metadata storage model, and outputting a storage result.

The invention provides a model compiling method, electronic equipment, a storage medium and a program product, relates to the field of artificial intelligence, and can solve the problem of low model calling efficiency in related technologies. The method comprises the steps of obtaining compiling information corresponding to a target model from a database according to structure information of the target model; the database is used for storing structure information of the model and corresponding compiling information; and compiling the target model based on the compiling information. The model calling efficiency can be improved.

Aspects of the disclosure provide a method performed by a central Network Data Analytics Function (NWDAF) in a communications network. The communications network comprises one or more local NWDAFs configured to develop a model using federated learning, in which each local NWDAF stores a copy of the model and trains the copy of the model by inputting training data into a machine-learning process. The method comprises receiving, from the one or more local NWDAFs, a respective local model update comprising an update to values of one or more parameters of the model generated by training a respective copy of the model using machine-learning. The method further comprises combining the local model updates received from the one or more local NWDAFs to obtain a combined model update.

The platforms, systems and methods provided herein may provide a data-driven workflow platform. The method comprises: mapping selected data objects to a data storage model of the data-driven workflow platform, where the selected data objects are stored in a data cloud configuration that is operatively coupled to the data-driven workflow platform; and displaying a graphical user interface (GUI) for building a workflow that utilizes or manages the selected data objects, the workflow comprises at least one automated action triggered by a triggering event of the selected data objects and wherein the triggering event is a detected change of the selected data objects based at least in part on an output of a machine learning algorithm trained model.

Methods, systems, and computer-readable storage media for receiving, from probes, probe data indicative of gas storage, the probe data being collected by probes included in operating wells and observation wells within a field, the probe data comprising surface data and subterranean data indicative of a health of a gas storage reservoir within the field. A gas storage status is determined by using a gas storage model, the gas storage model correlating the surface data and the subterranean data within the field. A gas storage assessment report including a pressure map reflecting the gas storage status within the field is provided.

The invention relates to the technical field of information retrieval and processing, and discloses a cross-format document information traceability positioning method and system, a terminal and a medium. According to the method, document contents in different formats are decomposed into standardized element blocks and element blocks with position information, and a unified storage model is constructed, so that a corresponding file pre-conversion result is generated; based on the file pre-conversion result, vectorizing a to-be-positioned information extraction result, executing semantic similarity retrieval in a vector database to generate a candidate element block set, and on this basis, executing refined matching of text similarity and entity similarity in combination with candidate element blocks and associated element blocks, so as to obtain a file pre-conversion result; determining the accurate position of the information extraction result in the original document; according to the method, unified position sensing modeling of the cross-format document can be effectively realized, and a positioning blind area is eliminated.

A system for supporting artificial intelligence / machine learning (AI / ML) model functions using a service-based architecture in a radio access network (RAN) intelligent controller (RIC) is provided. The system includes a first function for managing AI / ML functions, and for exposing management and exposure services for the AI / ML functions. The system also includes a second function for providing services for deploying the AI / ML models in the at least one RIC, and a repository for storing the AI / ML models. The first function, the second function, and the repository are connected with the service-based architecture.

The invention discloses a dam multi-source data management method, system and device and a medium, and relates to the technical field of data processing, and the method comprises the steps: receiving monitoring data from a plurality of heterogeneous data sources, and carrying out the classification processing of the monitoring data; performing format conversion, identifier unification and data association processing on the received monitoring data; dividing the standardized data into a shared type resource set, and performing data interaction according to a preset exchange mode; dividing the data according to the type, the structural feature and the access requirement of the monitoring data, and storing different types of data in a database; and according to the data type, a differentiated backup strategy and a storage model are adopted for management. According to the method, the organization efficiency and the query performance of mass monitoring data are improved by constructing the basic database, the monitoring database, the spatial database, the multimedia database and other multi-layer structures and combining hierarchical storage, index optimization and structural standardization table design, and the multi-dimensional data service requirement is met.

A device for editing large language models and multimodal AI models. The device includes a memory medium that stores the data and / or dataset of the model, a control unit that is configured to process the data within modules and according to requests and pre-defined configuration parameters from a user and / or client which communicates with the device with at least one communication unit to provide at least one output based on an input data; and runs the instruction on the trained AI model. The device includes a compression module, a lobotomization module and an inception module and at least one of the modules is configured to identify layers of the trained model with the weight matrices; and decomposing the weight matrices of the trained model into a tensor network structure. The device also includes an integration module that provides connection and data transfer between external systems and platforms and the device.

Proposed is a calibration method for a lidar and an IMU using features according to the shape of a specific object included in point cloud data. The method may include placing point cloud data acquired by a lidar mounted on a vehicle on a predefined world coordinate system, by a data generator, extracting a region to be used for calibration from the placed point cloud data, by the data generator, identifying at least one object included in the extracted region, by the data generator, and performing calibration on the point cloud data by fitting point cloud included in the at least one identified object to a pre-stored model, by the data generator. The present method is a technology developed with support from the Ministry of Land / KAIA (Project No. RS-2021-KA160637).