30results about How to "Resolving heterogeneity" patented technology

The invention relates to a spatial information network architecture. The spatial information network architecture comprises a ground information port, a data forwarding layer and a space-based backbone network; the ground information port is connected with the data forwarding layer, is connected with the space-based backbone network through the data forwarding layer and is used for executing a user task request according to a received flow table; the data forwarding layer is composed of at least one satellite selected from a medium earth orbit satellite and a low earth orbit satellite and is used for receiving the flow table, calling resources to execute the user task request, performing data transmission and forwarding according to the flow table; the space-based backbone network comprises at least one geostationary orbit space satellite, a controller in the geostationary orbit space satellite allocating resources for the task request and issuing the flow table; and the ground information port and the space-based backbone network together form a double-backbone control system. According to the spatial information network architecture of the invention, the global coverage characteristics of the GEO (geostationary orbit) satellite and the computing power of a ground base station are effectively combined, and the spatial information network can obtain route forwarding strategies in real time, allocate resources and nodes reasonably according to tasks and improve resource utilization rate and node cooperation ability.

The invention provides a hierarchical federated learning method and device based on asynchronous communication, terminal equipment and a storage medium, and relates to the technical field of wirelesscommunication networks. The method comprises the following steps: an edge server issues a global model to an intra-cluster client to which the global model belongs; the client updates the model by using the local data and uploads the model to each belonging cluster edge server; the edge server determines to update the clients in the cluster according to the client update uploading time; the received model parameters are averaged by the edge server, and it is selected to asynchronously upload the model parameters to the central server or directly issue the model parameters to the client according to the updating times of the current client; and the central server performs weighted averaging on the parameters uploaded by the edge server, and issues the parameters to the client for training until the local model converges or reaches an expected standard. According to the method, the federated learning task can be efficiently executed, the communication cost required by the parameters of the federated learning model is reduced, the edge server butted with the client is dynamically selected, and the overall training efficiency of the federated learning is improved.

The invention provides an automatic and flexible method of virtual machine live migration in the field of cloud computing. The method can deal with severe network environments, reduce system resource consumption, improve system usability, and obtain ideal system stop time. According to the memory transfer scheme of the method, memory page pre-copy technology is adopted so as to achieve the purpose of reducing stop time, and a central control engine is used for real-time monitoring of utilization conditions of resources in a cloud computing environment, the live migration can be initiated independently. Meanwhile, the method realizes the technology of live migration between virtual machine monitors which are different in type, and increases flexibility of the live migration.

The invention belongs to the technical field of cost analysis of building engineering, Disclosed are an analyzer for construction cost and a control method thereof. The analyzer for construction costcomprises a data acquisition module, a data extraction integration module, a main control module, a cost list classification module, a data analysis and processing module, a power network cost analysis module, an audit module, a data conversion demonstration module and a data application sharing module, wherein, the analyzer comprises a data acquisition module, a data extraction integration module, a main control module, a cost list classification module, a data analysis processing module, a power network cost analysis module, an audit module, a data conversion demonstration module and a dataapplication sharing module. The cost list classification module of the invention is convenient to summarize the existing problems in an all-round way, and prevents frequent changing rules and adding new constraints to the original rules in actual operation, thereby saving time and labor. At the same time, through the cost analysis module of power network to solve the heterogeneity of big data of different business, truly realize the information intercommunication of big data system, comprehensive and efficient operation; In addition, the audit module can ensure the audit quality and reliability by auditing data, auditing and audit results table confirmation and other components of the multi-node multi-dimensional audit process.

A multi-spectral-coverage synthetic environment simulation system comprises a synthetic environment modeling and calculating unit, a synthetic environment data Sedris standardization unit and a synthetic environment rendering and visualization unit. The defects in the prior art are overcome, and the problems that in a simulation test, the complex and true environment is difficult to simulate, and meanwhile accurate and effective environment support data are insufficient, so that influences on the test by environmental factors are not sufficiently embodied are solved. modeling, SEDRIS standardization and visualization of the synthetic environment of atmosphere, infrared, electromagnetism and others are achieved, standardized synthetic environment data are provided for simulation application, and support is provided for operational effectiveness evaluation of a weapon system.

The invention discloses a map reduce computing method under an internet environment. The method utilizes available computation and storage resources of an internet desktop personal computer (PC) to carry out map reduce computation. The desktop PC is added to a system in a volunteer form and forms a work node. A user as a master uploads to-be-processed data to a server. The server schedules map and reduce tasks to the work node for execution. When all the tasks are finished, final results are fed back to the user automatically. On the basis of metadata management, a data driving way is adopted, task scheduling is converted into data scheduling, the problems that node dynamic fails easily and influence is caused are solved, and therefore a fault-tolerant map reduce computing model and the method are simple and effective under the internet environment. Due to the method, an existing desktop PC is utilized by campuses, enterprises and governments to construct a map reduce computation platform with private purposes for carrying out mass data analysis, or the method is used for improving a volunteer computing system to support the map reduce computing model and lowering difficulty of application development of a volunteer commutating system.

The invention relates to a monitoring video multi-dimensional pedestrian behavior intelligent analysis system and method. The system comprises an individual behavior analysis module, a small group behavior analysis module, a group behavior analysis module, a scene analysis module, a pedestrian behavior dimension evaluation module, an intelligent gateway, an application logic integration module anda video source management module. The method comprises the following steps: acquiring a monitoring video data stream, extracting video frames from the monitoring video data stream, and performing scene classification and pedestrian behavior dimension selection by using key video frame data; combining the scene and pedestrian behavior dimension information; performing decisions selected by the pedestrian behavior analysis module and transmission of management data; and performing intelligent behavior recognition through a pedestrian behavior analysis model of a scene integrated in the pedestrian behavior analysis module under the selected behavior dimension, and outputting a final pedestrian behavior recognition result. Pedestrian behavior recognition in various complex monitoring scenes and pedestrian behavior dimensions can be supported.

The invention discloses a method and a system for incremental machine learning based on transparent computing. An incremental machine learning framework based on transparent computing is used, so thatthe system can perform machine learning on the premise of low network dependence. Training data is split and processed, so that time consumption of machine learning is greatly reduced on the premiseof ensuring test accuracy, and machine learning efficiency is improved. In addition, through incremental feedback learning, accuracy of machine learning is continuously improved.

The invention discloses an IBN-based Internet-of-things satellite system and a routing method thereof. The system includes an application layer, which is used for submitting intents of user requests to a control layer; the control layer, which is used for receiving service intents of the application layer, and translating the service intents of the application layer into implementable solutions toissue the same to a data layer; a data layer, which is used for data forwarding; and a maintenance layer, which includes a sensing module and a self-healing module. The sensing module is used for sensing network status according to status information and traffic information on the data layer and all other network elements, and maintaining smoothness of a network according to link load and trafficsizes; and the self-healing module is used for detecting network anomaly situations, establishing firewalls, and predicting security problems. The system ensures safe and automatic maintenance in anInternet-of-things satellite, also solves heterogeneity problems of Internet-of-things equipment, and effectively reduces probability that network configurations cause the network to be not smooth.

The invention discloses an agent-based business span-domain operation supporting system and a method. A locally-running operation supporting system agent of an allochthonous operator is utilized to authorization and measurement of business. When a client uses the allochthonous business, a local authorization system requests the locally-running agent for authorization information and issues the authorization information to the client; and two operator bosses can accurately record the usage amount of the allochthonous business according to the issuing number of the authorization information. Themethod not only can separate concrete certificate authority methods of two parties, but also can ensure the usage amount consistency of the business, and meanwhile, synchronization of client data andorder information between the two bosses are avoided, thereby being beneficial to protecting the local client data. The amount of the allochthonous business is processed by the agent and sent to allochthonous bosses in fixed time, and remote communication adopts an asynchronous mode to reduce the transinformation amount of remote information and lessen the dependence on the reliability of a remote network.

The invention discloses a malicious social user detection method based on adversarial learning. The malicious social user detection method mainly solves the problems that an existing malicious social user detection method is low in precision and poor in stability and reliability. According to the scheme, the method comprises the steps of 1) collecting data and establishing different databases; 2) constructing different feature maps according to the association between the user and the databases; 3) calculating different feature preferences according to the feature maps corresponding to the social data nodes of all published content of the user; 4) generating a training sample set and a test sample set; 5) constructing a neural network; 6) taking different feature preferences of the training samples as input, and training the neural network; and 7) inputting a test sample into the trained neural network to obtain a detection result. According to the method, the content feature preference, the attention feature preference and the propagation feature preference are comprehensively considered, the detection precision, stability and reliability are improved, and the method can be used for public opinion analysis, community discovery, false news detection and social network marketing.

A universal virtual experiment real object equipment access system comprises a hardware interface protocol converter, a virtual experiment proxy node and a configurable automatic real object equipment access gateway generator. The hardware interface protocol converter comprises a protocol conversion unit and a bus interface. The universal virtual experiment real object equipment access system has the advantages that on the basis of user software interface configuration, a virtual experiment access gateway matched with a real object equipment transmission protocol is automatically generated, the heterogeneity problem of the transmission protocols and the transmission interface is solved, and rapid and seamless linkage of real object equipment and virtual equipment middleware is achieved. According to the universal virtual experiment real object equipment access system, on the basis of flexible protocol configuration, rapid and seamless integration of a real object virtual experiment verification platform is achieved, the heterogeneity problem of the various transmission protocols is solved, user zero-code compiling is achieved, and the development period of the virtual experiment verification application is shortened.

The invention belongs to the technical field of intelligent equipment user behavior recognition, and particularly relates to a user behavior recognition method based on personalized semi-supervised online federated learning. The method is characterized by comprising the following steps: determining a client with a label and a client without a label; performing online semi-supervised federated learning by a FedHAR algorithm, and training a generalized neural network model; and performing personalized federal fine tuning on the generalized neural network model to obtain a multi-modal personalized neural network model. A semi-supervised online learning and federated learning framework is used, and a personalized user behavior recognition method is provided to solve the behavior recognition problem in a real scene and the privacy problem.

The invention provides a cognitive method and system based on a fused multi-modal data matrix, and the method comprises the steps: obtaining a target multi-modal data set, obtaining a cognitive target task corresponding to the target multi-modal data set, and taking the cognitive target task as a to-be-cognitive target task; performing feature extraction on the multi-modal data set to obtain a feature representation of each piece of single-modal data, and performing vectorization processing on the feature representation of each piece of single-modal data to obtain a word vector matrix fused with multi-modal features; and establishing a semantic network expression cognitive model, and calculating the similarity between the features in the word vector matrix fusing the multi-modal features and the to-be-cognitive target task to obtain target feature knowledge corresponding to the to-be-cognitive target task. According to the method, the problem that an existing information processing mode cannot meet the requirement for rapidly recognizing semantic information of different sources is effectively solved, and structured arrangement and fusion cognitive calculation of the multi-modal data set are achieved.