2083 results about "Subnetwork" patented technology

The invention provides a micro-grid energy management method and system based on deep reinforcement learning, and relates to the technical field of power grids, and the method comprises the steps: constructing a dual-time scale deep reinforcement learning model which comprises a day-ahead scheduling sub-network and a real-time scheduling sub-network; the day-ahead scheduling sub-network predicts micro-grid operation strategies at a plurality of time points in the future based on the long short-term memory network; the real-time scheduling sub-network is based on a depth deterministic strategy gradient algorithm, real-time state data and a day-ahead scheduling prediction result are fused to construct an evaluation function, an instant reward value is calculated, and renewable energy power generation, energy storage charging and discharging and an external power grid electricity purchasing and selling power adjustment instruction are optimized online. According to the invention, through dual-time-scale collaborative optimization, the energy management efficiency and economic benefits of the micro-grid are improved, and the operation stability of the micro-grid is enhanced.

A method for establishing connections and forming groups in an edge computing system includes detecting and identifying devices attempting to connect to the network using a processor. The method involves authenticating detected devices with a common pre-shared key (PSK) stored in memory, forming groups of connected devices based on predefined criteria, and sharing the PSK within each group via a secured channel. It also includes creating a subnetwork or private LAN for each subscriber using network configuration data, assigning virtual pre-shared keys (vPSKs) to devices based on service requirements, determining device capabilities by analyzing received device-specific information, and identifying supported applications based on device capabilities and application compatibility data stored in memory.

The invention belongs to the technical field of communication transmission, and discloses a PLC networking data acquisition system and method based on an edge gateway, and the system comprises a dynamic virtual isolation module which is used for identifying and classifying different PLC devices through the deployment of an SDN controller and the adoption of a deep packet inspection technology, automatically distributing the different PLC devices to a virtual subnet, and obtaining a virtual IP; each virtual subnet isolates the PLC equipment with the same IP address through a virtual switch and a virtual router; recording the virtual IP and actual physical network port information of the PLC equipment through a mapping table, and establishing a network topology view; the multi-protocol analysis module is used for connecting the edge gateway and the PLC equipment based on a network topology view, carrying out protocol analysis by integrating a PLC communication protocol library and adopting FPGA hardware acceleration, and reading PLC networking data from the PLC equipment in real time; and the flexibility and the expandability of data acquisition can be improved.

The invention relates to the technical field of system security, in particular to a terminal zero-trust security capability system based on trusted environment perception. According to the invention, through real-time environment perception, dynamic risk assessment and fine-grained access control, all-around security protection is provided for enterprises; firstly, an identity agent program is deployed on a user terminal, and environment risk information is collected and reported in real time; secondly, a secure access control area is set up and comprises a trusted environment perception system, an intelligent identity analysis platform, a trusted access console and other core components for risk assessment, the trust level is obtained through user behavior analysis, and dynamic authorization is conducted based on the risk and the trust level; and finally, realizing fine-grained access control on applications and interfaces in the data subnet through a trusted application agent and a trusted interface agent. Environment perception, identity analysis and access control are closely integrated to form a coherent zero-trust architecture, environment changes and threats can be better responded, and the system security is improved.

The invention relates to the technical field of traffic control, in particular to a traffic signal cooperative control method based on multi-agent reinforcement learning, and the method comprises the following steps: modeling each signal lamp intersection in a road network as an agent, and constructing a distributed multi-agent control environment; dividing the whole road network into a plurality of small subnets according to spatial correlation, and sharing and aggregating traffic information in the subnets through a neighborhood information sharing mechanism; utilizing a space-time diagram attention network to extract traffic state characteristics including a space-time dependency relationship in an intersection agent and a neighborhood thereof; and defining a traffic state, a traffic signal control strategy, a reward and punishment function, a network architecture and a target function required by training of the distributed intelligent agent, and carrying out joint training on the distributed intelligent agent until a training target is achieved. According to the invention, the real-time sensing capability of the signal control intelligent agent to the traffic flow dynamic state can be improved, and the collaborative decision-making capability among multiple intelligent agents is enhanced.

The invention discloses a defect diagnosis method and system based on multi-modal data cooperative training, and belongs to the technical field of defect diagnosis, and the method specifically comprises the steps: constructing a multi-modal cooperative diagnosis model comprising a feature extraction sub-network and a cross-modal attention module; after multi-modal data is collected and preprocessed, initial features are obtained through the feature extraction sub-network, attention weights are generated through the cross-modal attention module, weighted multi-modal features are obtained, multi-scale fusion features are obtained through the multi-scale feature extraction sub-network, and a preliminary diagnosis result is given through cascade processing. Meanwhile, the data integrity is detected, and a modal missing scene is coped with through cascade collaborative diagnosis; and finally, comparing the two types of diagnosis results with a defect labeling sample to obtain a multi-modal collaborative diagnosis model after training optimization, inputting to-be-diagnosed sample data, outputting a final diagnosis result and updating the defect labeling sample, and realizing efficient and accurate defect diagnosis.

The invention discloses an SAM2 multi-task perception binary segmentation method based on a hybrid expert adapter. The method is specifically implemented according to the following steps: step 1, constructing a data set and an encoder; step 2, constructing a hybrid expert adapter module; and step 3, constructing a task awareness gating module. According to the method, a pre-trained SAM2 is taken as a main network, on the basis of freezing the main body weight, a standard adapter and a MoE-Adapter are respectively deployed on odd and even layers of an encoder, and a lightweight expert sub-network and a dynamic gating strategy are combined, so that unified processing of multiple tasks such as salient target detection, camouflage target identification, marine animal segmentation and the like is realized.

The invention discloses a route configuration method and device, equipment, a storage medium and a product, and the method comprises the steps: directly configuring a route prefix on network equipment or declaring the route prefix through a sub-network in a user sub-network multi-homing access scene, enabling the network equipment to update a source address verification rule or add a corresponding static route after receiving route configuration information, and enabling the network equipment to access to a user. And the network equipment generates a corresponding source address verification rule after receiving the declaration information, so that accurate protection of more scenes such as asymmetric routing, multi-homing asymmetric access of intra-domain users, traffic bypass and the like can be realized, and the problem of mistakenly filtering legal traffic is avoided.

The invention provides a knowledge graph-based security event automatic response method, which comprises the following steps of: acquiring multi-source security data, performing de-duplication and standardization processing, and generating a structured security data set; based on the network security ontology model, entities and relationships are extracted from the data set, and a security knowledge graph is constructed through entity alignment and conflict resolution; according to ATTamp; the CK framework divides network subnets, generates attack path diagrams of the subnets, and fuses the attack path diagrams into a global attack graph. Then, abnormal behavior nodes in the security knowledge graph are analyzed, and risk scores of the service layer, the host layer and the system layer are calculated in combination with the global attack graph; and matching a predefined response strategy library based on the risk scores, executing operations such as banning an IP (Internet Protocol), isolating a host or updating firewall rules and the like through an SOAR platform, generating a disposal report and updating a security knowledge graph. According to the invention, the network security event processing efficiency and accuracy can be improved, and the overall security protection capability of the system is enhanced.

The invention relates to the field of remote sensing image processing, and discloses a high-quality HRHS image generation method based on a double-subnetwork structure, and the method comprises the following steps: receiving an input low-resolution hyperspectral image and a panchromatic image, which are respectively used for providing spectral information and spatial information; performing noise filtering on the low-resolution hyperspectral image through an anti-noise pre-fusion sub-network, and generating a pre-fused hyperspectral image by using the correlation between the high spatial resolution information of the panchromatic image and the spectral characteristics of the low-resolution hyperspectral image; performing high-frequency detail reconstruction and enhancement on the pre-fused image based on a conditional diffusion model through a high-frequency enhancement sub-network; and fusing the low-frequency information of the pre-fused image and the high-frequency detail information generated by the high-frequency enhancement sub-network through wavelet inverse transformation, and outputting a final high-resolution hyperspectral image. According to the method, spectrum and space information can be effectively fused in various noise environments, and a high-resolution hyperspectral image with hyperspectral fidelity and space details is generated.

The invention belongs to the field of unmanned aerial vehicle communication, and discloses a flight ad hoc network distributed routing method based on multi-agent reinforcement learning, and the method comprises the steps: building four network performance models: energy consumption, packet delivery rate, throughput and end-to-end time delay; establishing a node interaction mechanism; establishing a Markov decision process, defining definitions of a state space, an action space and a space state transition probability, and designing a reward function; based on an improved multi-agent depth deterministic policy gradient algorithm DE-MADDPG, designing a flight ad hoc network distributed routing method based on multi-agent reinforcement learning; each unmanned aerial vehicle modifies a running protocol and a protocol parameter according to local network information of each node, optimizes a main network and a subnetwork of the system, and improves the overall performance of the network. The method mainly solves the problems that a traditional flight ad hoc network protocol is not high in adaptive capacity, and the network environment is not matched with the protocol. Compared with a traditional method, the method has the advantages of being low in control delay, high in control precision and good in performance.

The invention relates to the technical field of data processing, and provides a network device discovery method, which can perform iterative retrieval under a target network according to an ARP table, provides a more comprehensive view, and is beneficial to monitoring the health condition and the network load condition of the target network. Legality detection of an IP address is carried out based on a rule base; legality detection of an MAC address is carried out based on a multi-level detection mechanism; uniqueness detection is carried out based on a device fingerprint; subnet suitability detection is carried out based on a subnet mask matching algorithm; a self-adaptive network equipment discovery algorithm combined with multi-dimensional data analysis can reduce redundant data, eliminate invalid information and improve the accuracy and management efficiency of equipment discovery; information generation is carried out based on the initial network equipment data to obtain the network equipment view, so that more accurate network equipment discovery is carried out.

Disclosed are apparatuses, systems, and techniques that may use machine learning for reversible translations of speech utterances. The techniques include training and using duplex neural networks (NNs) having a first subnetwork and a second subnetwork that are mirror images of each other. Training data for training the duplex NNs may include a target output that includes a first speech utterance in a first language, a first training input that includes the target output distorted by a noise, and a second training input that includes a second speech utterance in a second language. The duplex NNs may be trained to identify, using the first training input and the second training input, at least one of the target output or the first noise.

A system comprises an encoder configured to compress and encode data for a three-dimensional dynamic mesh. The three-dimensional dynamic mesh is compressed and signaled using a base-mesh sub-bitstream and an atlas sub-bitstream (e.g., displacement and / or attribute sub-bitstream). The atlas sub-bitstream comprises displacement information to be applied at sub-division points of the base mesh to recreate the three-dimensional dynamic mesh at a given moment in time and may also include attribute values, such as colors, textures, etc. to be applied to the reconstructed mesh. Sub-mesh identifiers are included in tile headers of an atlas sub-bitstream that enable selective reconstruction of sub-meshes. A complimentary decoder decodes a compressed bitstream for the three-dimensional dynamic mesh, encoded such as by the encoder.

The invention belongs to the field of fault monitoring, and particularly relates to a fault self-healing method and system under an intelligent operation and maintenance global P2P architecture, and the method comprises the steps: constructing a global operation and maintenance monitoring double-layer node network which comprises a basic monitoring subnet and a collaborative management and control subnet; the basic monitoring subnet collects platform operation data in real time according to a preset inspection strategy, determines an abnormal root node and an association node through causal association analysis, generates an early warning signal and uploads the early warning signal to the collaborative management and control subnet; after the collaborative management and control subnet receives the early warning, a fault self-recovery strategy is started, resources are dynamically allocated, abnormal node data batch migration and recovery are executed, and a self-recovery result is fed back to the basic monitoring subnet to trigger loop check until abnormity is eliminated; according to the method, real-time monitoring and intelligent analysis of operation and maintenance data are realized through cooperative work of the double-layer node network, and disaster tolerance and load adaptive adjustment capability of the system are effectively improved in combination with an automatic fault self-healing mechanism and data difference migration.

A method performed by an optical network controller includes: obtaining a topology structure of a subnet of a cluster, where the topology structure records an address of compute nodes in the subnet; determining a wavelength division configuration based on the topology structure, where the wavelength division configuration is used to allocate different wavelengths to the compute nodes in the subnet; providing the wavelength division configuration for the optical network switching device. Then the optical network switching device establishes an optical path between the optical network switching device and the compute node in the subnet based on the wavelength division configuration.

An artificial intelligence-based image processing system comprises a processor that executes instructions stored on a memory to classify an input image with a prototypical part neural network including a backbone subnetwork, a prototype subnetwork, and a readout subnetwork to produce an interpretable classification of the input image including one or a combination of a classification result of the input image and an interpretation of the classification result. The backbone subnetwork is trained with machine learning to process the input image with an incomplete sequence of active convolutional layers producing feature embeddings representing features extracted from pixels of different regions of the input image. The prototype subnetwork is trained to compare the feature embeddings with prototypical feature embeddings to produce results of comparison and the readout subnetwork is configured to analyze the results of comparison to produce the interpretable classification of the input image.

The invention discloses a power grid cross-correlation harmonic suppression method, device and equipment and a storage medium, and the method comprises the steps: carrying out the wavelet packet decomposition of a subnet electric signal, and obtaining the harmonic component of a subnet at each frequency band; calculating the harmonic phase difference between the subnets according to the cross-correlation function between the harmonic components of the subnets and the topological impedance matrix, and calculating the total harmonic distortion rate according to the harmonic components of the subnets in the frequency bands; constructing a harmonic transfer equation according to the harmonic phase difference between the subnets and the equivalent output impedance of the inverter so as to calculate the harmonic responsibility proportion of the subnets; solving the double-layer game model based on the total harmonic distortion rate, the reference value of the total harmonic distortion rate and the harmonic responsibility proportion of the subnet, and obtaining the output of the governance equipment; and inputting the output of the treatment equipment into the fuzzy controller to obtain the impedance adjusting quantity of the subnet, and adjusting the virtual impedance according to the impedance adjusting quantity to update the harmonic backflow path. According to the invention, harmonic suppression among multiple subgrids during high-permeability distributed grid connection is realized, and the operation stability of the power grid is improved.

The invention discloses a lightweight natural language processing large model training method. Comprising the following steps: obtaining processing language data; a labeled enhanced data set is obtained; a sparse mechanism of the sub-network is dynamically activated through semantic complexity, and an activation mask is obtained through calculation; generating a quantization bit width by activating the parameter sensitivity of the mask; performing cross feedback adjustment on the mixing precision quantification strategy; and evaluating the trained student model. According to the method, through dynamic sparse activation, mixing precision quantification and collaborative optimization, the core problems of static strategy stiffness, quantification precision loss, low knowledge migration efficiency and the like in large model lightweight are solved; dynamic sparse activation replaces traditional static pruning, and semantic loss is reduced; in order to realize optimized feature extraction of the enhanced data set, fusion optimization features are adopted; the mixing precision quantification effectively reduces the complexity of calculation.

Methods, systems, and apparatuses, including computer programs encoded on computer storage media, for training a denoising neural network and, once the denoising neural network is trained, generating new images using the denoising neural network. In particular, the described techniques include obtaining data specifying a trained initial denoising neural network, obtaining training data, and training, using the training data and on a denoising objective, a target denoising neural network. The target denoising neural network includes the core-subnetwork of the trained initial denoising neural network, but also includes a target encoder and decoder neural networks that are not included in the trained initial denoising neural network. By training a target neural network that includes the core-subnetwork of a trained initial denoising neural network, the system achieves stable training of a large-scale target denoising neural networks that can be used to generate high resolution images.

The invention relates to the technical field related to automatic driving, in particular to a trajectory prediction method based on adaptive interaction and dynamic intention, which comprises the following steps: firstly, constructing a heterogeneous interaction map, dividing a traffic scene into a vehicle grid, an environment grid and a non-driving area grid, and embedding multi-dimensional dynamic features; then dynamically adjusting a region of interest based on the behavior intention of the target vehicle, and extracting a high-correlation interaction subnet; modeling an interaction relationship by adopting a heterogeneous graph convolutional network, and processing the motion characteristics of the target vehicle and the neighbor vehicle through a sub-channel coding strategy; further realizing dynamic intention perception through a double-branch parallel attention architecture, and fusing macroscopic intention and dynamic intention information; and finally, iteratively generating a future trajectory prediction result based on a decoding architecture of a message passing mechanism. The method can effectively improve the long-term prediction performance in a lane changing scene, adaptively captures a dynamic interaction relationship, and improves the adaptability of a prediction system to the behavior intention change of a driver.

A computer system connected within a local network ingests telemetry data received from devices or sensors located within a premises and connected to the network. A virtual measuring instrument module instantiates within the computer system and includes communication protocol interfaces and predefined data transformation tools provisioned for interaction with telemetry data generated by the devices or sensors. The application associates with a subnetwork of the local network to which the devices or sensors are collectively connected. The subnetwork uses a first communication protocol that is different a protocol used on the local network. The communication protocol interfaces include a first interface that uses the first communication protocol. The telemetry data is received within the application from the devices or sensors on the first subnetwork using the first communication protocol interface as received telemetry data. The predefined data transformation tools translate a format of the received telemetry data to a normalized format. A presentation of the normalized telemetry data is configured for output to a user device.

The invention discloses an industrial ERP dynamic scheduling optimization method based on deep reinforcement learning, and the method comprises the following steps: S1, obtaining to-be-scheduled industrial ERP data, and carrying out the preprocessing of the to-be-scheduled industrial ERP data; s2, constructing a heterogeneous graph structure, and setting node types and edge weight parameters in the graph; s3, performing multi-layer graph feature extraction on the initial scheduling state graph based on a graph attention mechanism to generate a node feature vector; s4, constructing a distributed reinforcement learning environment, and setting an independent strategy sub-network for each type of scheduling sub-tasks; s5, performing joint decision-making on the scheduling action to generate a scheduling scheme as a current round scheduling result; s6, calculating a reward value and feeding back the reward value to the corresponding strategy sub-network; and S7, repeatedly executing the steps S1 to S6, and generating a new scheduling action by using the updated strategy sub-network. According to the method, adaptive optimization of industrial ERP scheduling under multi-source heterogeneous data is realized, and the resource utilization rate and the order delivery efficiency are remarkably improved.

The disclosure provides a method for assigning containerized workloads to isolated network constructs within a networking environment associated with a container-based cluster. The method generally includes receiving, at the container-based cluster, a subnet port custom resource specification to initiate creation of a subnet port object to assign a node to a subnet within the networking environment, wherein one or more containerized workloads are running on the node, in response to receiving the subnet port custom resource specification, creating the subnet port object, and modifying a state of the container-based cluster to match a first intended state of the container-based cluster at least specified in the subnet port object, wherein modifying the state comprises assigning the node to the subnet in the networking environment.

The invention provides a YOLO target detection method and system based on differential architecture search. The method comprises the following steps: firstly, constructing a candidate operation set; then, based on a YOLO network architecture, a super network is designed; then, pre-training the super network by using a differentiable architecture search method, and alternately optimizing network weight and architecture parameters; and finally, obtaining the searched subnet according to the architecture parameters, and introducing a coordinate attention mechanism to further improve the model performance. Different from a manual design method, the method has the advantages that by automatically searching the neural network architecture, tests and experiences required by manual work are greatly reduced, higher search efficiency and performance are achieved, and an efficient YOLO improved model suitable for target detection can be found more quickly.

The invention relates to the technical field of network space security, and provides a DoH malicious tunnel traffic detection method and system based on feature fusion, and the method comprises the steps: carrying out the preprocessing of obtained to-be-detected DoH traffic data, carrying out the sequence segmentation processing, obtaining a Token sequence, and extracting features based on a byte sequence feature extractor; statistical features are extracted and standardized, features are extracted through a statistical feature extraction sub-network, and statistical feature vectors are obtained; fusing the byte sequence feature vector with the statistical feature vector to obtain a classification result; the byte sequence feature extractor and the statistical feature extraction sub-network extract features, and training is carried out by adopting a semi-supervised learning framework of dynamic pseudo-tag screening for non-tag training samples. According to the method, multi-modal features are fused, a semi-supervised learning mechanism is introduced, malicious DoH traffic generated by multiple DNS tunnel tools is effectively identified under a limited annotation data set, and the detection precision and generalization ability are improved.

An example network provisioning system is configured to obtain one or more network slice subnet templates (NSSTs), including a template for a satellite access network (SAN) and, optionally, a template for a terrestrial access network (TAN). The network provisioning system defines a network slice template (NST) based on the one or more NSSTs. The network provisioning system deploys a network slice in accordance with the NST to enable service management and orchestration of the network slice including the SAN. The network provisioning system is configured to execute a user interface configured to display a representation of the network slice comprising the SAN. The network provisioning system is configured to receive via the user interface an indication to add a base station provided by a satellite and modify the network slice based in part on the selection. In some examples, the network slice is a 5G communications network slice.

The present disclosure discloses a data processing method and apparatus for a multi-blockchain, a computer device, a computer-readable storage medium, and a computer program product. The method includes: obtaining chain height voting information of a management node of a target chain for a target service subchain; determining, when determining, based on a subchain control contract on the target chain, that the management node is a target registered node, a target chain height of the target service subchain based on the chain height voting information and the subchain control contract; and generating height confirmation information corresponding to the target chain height based on the subchain control contract, and transmitting the height confirmation information to a verification node in the target service subnetwork, to cause the verification node to write the target chain height into a subchain management contract on the target service subchain.

The invention provides a method for efficiently detecting a fish target in a complex underwater environment based on a priori knowledge guide network, which comprises the following steps of: 1, acquiring an underwater image and preprocessing the underwater image; 2, establishing a recovery subnet module, a relation reasoning attention module and a self-adaptive feature fusion module, and performing complex underwater environment fish target detection; the recovery subnet module is used for guiding network learning to remove underwater turbid features through an underwater scattering model, generating a clear image through a water body recovery decoder WR, and reconstructing an underwater turbid image; the relation reasoning attention module is used for constructing a co-occurrence relation graph, performing relation reasoning by using a graph convolution network, and dynamically adjusting attention weight; and the adaptive feature fusion module is used for optimizing feature expression in combination with a Sigmoid function and a channel-by-channel weighting mechanism. The method can effectively cope with the conditions of sudden change of turbidity of a water body or complex background and the like, and can keep higher detection performance under different underwater conditions.

The embodiment of the invention relates to the technical field of wireless communication, in particular to a low-overhead multipoint time-frequency positioning method based on a federated learning framework, which comprises the following steps: establishing a broadband cognitive network system model, and selecting a time-frequency positioning model; the central server initializes network parameters of the time-frequency positioning model and issues the network parameters to each secondary user; a secondary user carries out local training on a local time-frequency positioning model based on a local data set, determines the weight of a subnet model, then extracts the gradient variation of each BN layer, carries out sorting based on the gradient variation, carries out dynamic pruning on the local time-frequency positioning model according to a target pruning rate, obtains a model mask, and uploads the model mask to a central server; and the central server complements the model by using each model mask, aggregates each subnet model weight to obtain a global weight, forms a personalized subnet model of each secondary user, and issues the personalized subnet model, thereby realizing low-overhead multi-point time-frequency positioning. According to the method, the communication overhead of the wireless link is effectively reduced, and the communication efficiency is greatly improved.