963 results about "Subnetwork" patented technology

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 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.

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.

The invention discloses systems and methods for privacy-preserving digital communications compliant with frameworks such as GDPR, DPDPA, HIPAA, and PSD2. A Virtual Identity (VI) is instantiated within a secure enclave and bound to one or more Compliance Jurisdiction Tokens (CJTs). Unlike conventional tokenisation limited to payment or aliasing, the invention enables multiple concurrent VIs (Multi-VI), each scoped to a declared purpose, jurisdiction, subnet, or session. Communication is permitted only if all bound CJTs validate inline, including Multi-CJT bindings where jurisdiction, purpose, and consent must all succeed, and Multi-Purpose CJTs (MCJTs) where multiple lawful purposes must be simultaneously satisfied. Each validation produces a Ledger-Anchored Validation Receipt (LAVR) containing pseudonymised metadata, anchored into tamper-evident ledgers for transparency. Regulators access receipts through a Regulator Query Interface (RQI), enabling filtered oversight without exposure of raw identifiers. The technical effect is to transform privacy and compliance into cryptographic enforcement across device, telecom, and network layers

The invention belongs to the field of artificial intelligence, particularly relates to a power plant infrastructure multi-target intelligent scheduling optimization method, and aims to solve the problems of static weight imbalance, disturbance response hysteresis and process coupling effect modeling insufficiency of traditional scheduling. According to the method, a multi-dimensional space-time semantic model is constructed, BIM, sensor and environment data are integrated, construction period, cost, resource and safety four-dimensional target weights are dynamically set, and an improved non-dominated sorting genetic algorithm is adopted to generate an initial Pareto optimal schedule; and then inferring an inter-process nonlinear coupling delay factor through a graph neural network, embedding a disturbance response module, starting local rolling re-optimization when a progress deviation or an external event is detected, limiting an influence subnet and freezing a stable region. According to the scheme, stage self-adaptive target focusing, chain risk pre-buffering and minute-level robust adjustment are achieved, the stability of a critical path is improved by 45%, the secondary optimization frequency is reduced by 60%, the calculation efficiency and the execution toughness are both considered, and efficient and accurate landing of a large power plant infrastructure project is supported.

A distributed cloud computing system is disclosed that includes a controller configured to deploy network constructs including any of transit gateways, spoke gateways, subnets, or private networks and logic that, upon execution by one or more processors, causes performance of operations including: causing rendering of a graphical user interface that includes a display panel configured to display progress of a build process for a network topology graph, receiving first user input through the graphical user interface indicating selection of a first cloud service provider, a first access account, and a first cloud region, receiving second user input through the graphical user interface indicating selection of one or more of the network constructs to be deployed in the first cloud region, instructing the controller to deploy the one or more of the network constructs in the first cloud region according to the first user input and the second user input.

The invention discloses a thermal power plant operation risk intelligent early warning method based on deep learning. The method comprises the following steps: S1, collecting operation data to construct a multi-modal time sequence; s2, respectively inputting modal sub-networks after standardization processing; s3, setting a reversible residual structure for each sub-network and introducing a dynamic scaling mechanism; s4, introducing a time gating mechanism to adjust residual response; s5, fusing the modal features to form fusion representation; s6, performing forward transformation, combining with intermediate state reverse reconstruction, and calculating a residual error; s7, a sliding window aggregates residual errors to generate a scoring sequence, and risk levels are output; and S8, generating a key factor sequence based on residual attribution analysis features. According to the invention, operation abnormity in-advance identification and risk source accurate positioning under complex working conditions are realized.

The invention discloses a generative adversarial network architecture search method and system and an image generation method, and belongs to the technical field of network architecture search. The searching method comprises the following steps: performing single-path sampling on a pre-constructed generator super network according to a parameter quantity constraint range to obtain an effective subnetwork; training the generator super-net by adopting a complexity adaptive learning rate optimization strategy to obtain a pre-trained generator super-net; adversarial training is carried out on the generator hypernet and the discriminator to obtain a pre-trained discriminator; generating a generator super-network candidate architecture through a genetic algorithm in the early stage of the evolution stage and through a covariance matrix self-adaptive evolution strategy in the later stage of the evolution stage; and performing multi-target non-dominated sorting on the generator super-network candidate architecture, updating an effective sub-network and keeping a Pareto optimal solution to obtain a searched optimal generator architecture and further obtain a searched optimal generative adversarial network architecture. The method not only ensures the search quality, but also improves the calculation efficiency.

The invention discloses a data processing method and device of a data center network, equipment, a medium and a product, and relates to the technical field of data processing, and the method comprises the steps: obtaining network topology information and a data forwarding table rule corresponding to each equipment subnet in the data center network; performing parallel coding on each piece of network topology information and each data forwarding table rule based on a network parallel binary decision diagram to obtain a data forwarding path structure between subnets on each piece of equipment, the network parallel binary decision diagram being generated based on a lockless index and hierarchical cache; based on the path connectivity corresponding to the data forwarding path structure, constructing a data forwarding graph corresponding to the subnet on each device; and verifying the data reachability between the subnets on each device according to the data forwarding graph, and generating a data reachability verification result. According to the technical scheme, the concurrency limitation is overcome, high concurrency can be achieved, and the data plane verification efficiency of the data center network is improved.

The invention relates to the technical field of network intercommunication, and provides a method and equipment for realizing a bare metal gateway based on an OpenFlow flow table, and the method comprises the steps: enabling a bare metal server to be accessed to a VPC network, and creating and binding a network port to a specified bare metal gateway node in a specified VPC subnet through a cloud management platform; a logic flow table is generated through the OVN controller according to the network ports and the binding relation of the network ports, and the logic flow table is issued to the bare metal net nodes bound to the network ports; the bare metal net node translates the received logic flow table into an OpenFlow flow table and issues the OpenFlow flow table to a local OVS virtual switch; and the OVS virtual switch performs two-layer forwarding processing on the message transmitted between the bare metal server and the virtual machine in the VPC according to the received OpenFlow flow table. The bare metal gateway is realized in a pure software mode, the dependence on special hardware is avoided, and the cost is remarkably reduced. The efficient conversion between the VXLAN and the VLAN is realized based on the OpenFlow flow table, the flexibility, the expandability and the operation and maintenance automation degree are improved, and the high-performance two-layer intercommunication is ensured.

The invention provides a large-scale constellation network semi-physical simulation test system and method, the system comprises a constellation simulator, a satellite-borne router, ground detection equipment and a protocol conversion network bridge, a Mininet simulation platform is deployed, and dynamic topology adjustment, an OSPF dynamic routing protocol and a multi-scene test function are supported; a user can observe network topology change, route updating time and key performance indexes in real time. A semi-physical simulation architecture is realized through unified IP subnet interconnection and the same routing protocol, a space-time synchronization engine ensures that virtual and real node routing tables take effect at an appointed moment, and in combination with a protocol consistency sandbox, hidden faults caused by inconsistent behaviors of hardware and software protocol stacks are solved, and the service life of the virtual and real node routing tables is prolonged. A plurality of real nodes are accessed in a large-scale network environment, and end-to-end data transmission test, security policy verification and protocol compatibility analysis are carried out; the invention provides an efficient, flexible and reliable solution for the design, verification and optimization of the satellite network, and has a wide application prospect.

The invention discloses a quantum security multi-party computing method of a hybrid topological structure, which is applied to a quantum security multi-party computing system, the quantum security multi-party computing system comprises m sub-networks of star network topological structures, and each sub-network corresponds to a quantum center and a plurality of users; quantum communication is completed between every two quantum centers based on a ring network topology structure; the quantum security multi-party computing method comprises the following steps: each quantum center completes modular addition operation of a corresponding user secret value in a corresponding subnet to obtain a modular addition result of all user secret values in the corresponding subnet, and the modular addition result is recorded as a subnet secret value modular addition result; and on the basis of a subnet secret value modular addition result, each quantum center continuously participates in the secure multi-party computing protocol among the quantum centers to obtain the sum of all user secret values in the quantum secure multi-party computing system. According to the method, on the premise that private secrets of users are not leaked, safer and more efficient quantum security multi-party calculation can be realized among the users with limited quantum capabilities.

The invention discloses a rail transit comprehensive dispatching emergency response method and system based on the Internet of Things, and relates to the technical field of the Internet of Things. The method comprises the following steps: acquiring a rail transit comprehensive scheduling instruction queue through a rail transit Internet of Things terminal; sending the instruction queue to an emergency analysis engine, performing emergency scoring, and obtaining a scoring index queue; an RS-485 subnet layer is built by a plurality of paths of RS-485 interfaces arranged by the gateway; the mapping relation between the instruction queue and the RS-485 subnets is established according to the scoring index queue, and the gateway responds to the instruction queue on the RS-485 subnets in a grading mode on the basis of the mapping relation. The technical problems that the emergency response is delayed and the communication efficiency is low due to the fact that an existing rail transit dispatching system cannot intelligently recognize the emergency degree of an instruction are solved, and the technical effects of improving the emergency response efficiency and optimizing network resource utilization through emergency scoring and RS-485 subnet grading response are achieved.

A system and method to protect resource servers from DDOS attacks is disclosed. A proxy gateway hides resource server using IP subnet addressing and receives authenticated IP packets. A cryptographic hash is generated, and destination IP packets address / ports are encrypted using a secret, algorithm, client IP address, resource server IP address, and port. The cryptographic hash and destination server are mapped into existing IP packet addresses, ports, or payload bits. The IP packets are routed to a redirect node in the public IP subnet of the proxy gateway. At the redirected node, the cryptographic hash is validated, and the resource server IP address and the port are extracted from mapping. Redirect node maps the clients public IP / port to an internal IP address and port associated with resource server and tunnels the IP packets from the redirect node to the resource server using the internal IP address and port.

The invention provides a training method of an index recommendation model and an index recommendation method and system.The training method comprises the steps that workloads and a candidate index set are obtained, and in the nth iteration, for each sub-network, recommendation indexes corresponding to the workloads are determined from the candidate index set so as to determine the local gradient of each sub-network, the index recommendation model comprises a plurality of sub-networks, updates parameters of the global network based on the local gradients corresponding to the sub-networks, and updates parameters corresponding to the sub-networks based on the updated parameters of the global network, the index recommendation model comprises the global network obtained through training convergence, and the index recommendation model is used for recommending index configuration for the to-be-recommended workloads. The training speed is greatly increased, and the stability and generalization ability of the index recommendation model can be improved.

The invention discloses a heavy-fork electrical system fault diagnosis and intelligent repair method, and particularly relates to the technical field of electrical system intelligent operation and maintaining.The method comprises the steps that multi-source data of electrical parameters, mechanical parameters and environmental parameters are collected through a sensor, and fault spatio-temporal feature extraction is conducted through a spatio-temporal attention feature extraction network; then obtaining a fault intelligent repair strategy based on the similarity between the real-time fault spatio-temporal feature vector and the historical fault spatio-temporal feature vector, issuing a corresponding intelligent repair instruction through a CAN bus, monitoring an execution result, evaluating execution result data after repair operation, and feeding back an evaluation abnormal result to a management terminal for human-computer interaction; according to the method, feature extraction is carried out by means of the network comprising the space attention sub-network and the cavity convolution time sequence feature extractor, the feature information of the system operation state can be comprehensively and accurately captured, and rich feature basis is provided for subsequent fault diagnosis and intelligent repair.

The invention discloses a knowledge graph enhanced retrieval method, system and device based on a large reasoning model and a medium, and the method comprises the steps: collecting documents processed by operation and maintenance personnel for establishing a knowledge graph database; performing network topology division on the knowledge graph in the knowledge graph database by adopting a grid division algorithm to obtain a plurality of sub-grids, and mapping a data set in the sub-grids into a sub-knowledge graph; inputting the question into the large model, selecting the sub-knowledge graph as a query range according to a keyword in the question, performing node expansion query and relation expansion query by the large model in the knowledge graph database, and returning an answer; matching the answers with the sub-knowledge maps, performing quadrant positioning on nodes of the sub-knowledge maps, and jointly generating a report by node coordinates, grid IDs and information associated with the answers; and calculating the text confusion degree of the report for measuring the performance of the large model, and if the text confusion degree is higher than a threshold value, adjusting the parameters of the large model. The knowledge retrieval speed is high, and the accuracy is high.

Techniques for structuring IP version 6 (IPv6) addresses into bit fields to embed language localization and services information (e.g., and / or other attributes / information) are disclosed. In some embodiments, a system / process / computer program product for structuring IPv6 addresses into bit fields to embed language localization and services information includes allocating a configurable portion of an IP version 6 (IPv6) address for a subnet, wherein the subnet includes information that is associated with a geographical location (geolocation); and determining an attribute for a new session based on the subnet.

The invention discloses a node separation learning method and system based on discrimination difficulty perception, and relates to the technical field of graph neural networks, and the method quantifies the discrimination difficulty through calculating a label distribution relaxation entropy value in a node multi-hop neighborhood; constructing a hybrid expert network architecture; splicing the entropy value and the node features and predicting a threshold value for a Top-p expert selection mechanism; activating the expert sub-network set according to a threshold value and weighting and combining outputs thereof to generate a node representation; and finally, obtaining a prediction result through a classifier and dynamically updating an entropy value in a training process. According to the method, computing resources can be adaptively allocated, and the accuracy and robustness of different configuration graph node classification are improved.

The invention belongs to the field of artificial intelligence, and provides a cloud AI model training method and a cloud AI system, and the method comprises the steps: collecting multi-modal state data and stimulation response data of a plurality of users in a process of using nerve electrical stimulation equipment; constructing a cross-user stimulation response evolution matrix; based on the stimulus response evolution matrix, extracting a plurality of individual feature subspaces and a shared response core subspace; based on the shared response core subspace, training a stimulation prediction main model, learning an average mapping relation between a state vector and feedback, and taking the average mapping relation as a shared basic model of each user model; configuring an individual regulation and control sub-network with a variable microstructure for each user, wherein the individual regulation and control sub-network is based on the shared basic model; and on the basis of the individual feature subspace, activating or freezing a part of submodules of the shared basic model to obtain the individual regulation and control subnetwork model.

Routing of packets in a partitioned subnet is realized in a data center (DC) by using a first type of address prefix (e.g., / 24 prefixes) to advertise respective hosts to its peers (i.e., external traffic). For routing traffic within the DC (i.e., internal traffic), a second type of address prefix (e.g., / 32 prefixes) that is longer than the first type. Switches / routers within the DC perform liveness probes detecting which hosts are connected thereto and updating a table (e.g., an ARP / ND table) to include the connected hosts (e.g., removing hosts that fail to respond). For changes to the table, a gate protocol (e.g., BGP or IGP) is triggered such that the switch / router advertises within the DC fabric the second type of address prefix for the host routes of the connected host. The switch / router also configures Liveness sessions for the connected hosts.

The invention discloses an urban road intelligent drainage scheduling system and method based on the Internet of Things. The method comprises the following steps that an urban road drainage Internet of Things sensing network is constructed, and a multi-source heterogeneous data set corresponding to an urban road target area is collected; performing space-time alignment and noise filtration processing on the multi-source heterogeneous data set, and meanwhile, performing ponding increase rate and pipe network pressure bearing evaluation to obtain urban ponding increase rates and drainage pipe network pressure bearing degrees corresponding to different urban road subnetworks; calculating drainage pipe network negative pressure indexes corresponding to the different urban road section sub-networks according to the urban water accumulation increasing rates and the drainage pipe network pressure bearing degrees corresponding to the different urban road section sub-networks; and acquiring an urban road network topological structure, performing drainage hierarchical scheduling processing, issuing the urban road network topological structure to drainage facilities in corresponding areas to execute drainage scheduling and feed back scheduled drainage effect data, and re-executing closed-loop scheduling if the drainage effect data is not accordant with the drainage effect data. According to the invention, accurate drainage scheduling of the urban road area can be realized.

The invention discloses a multi-domain combat dynamic deduction system based on deep reinforcement learning, and relates to the technical field of military automation, and the system comprises an environment perception-confrontation game-emergence mining three-layer nested dynamic architecture in which a bidirectional feedback mechanism, a cross activation mechanism and an adaptive adjustment node are deployed. The three-layer nested dynamic architecture comprises a dynamic adaptive adversarial network, a chaotic data weaving device and an emergence capture and evolution guider; according to the invention, through a three-layer nested architecture and core module design, multi-domain information real-time interaction and cross-level collaboration are realized, and the multi-domain linkage simulation precision is improved; by means of a heterogeneous subnet dual-mode interaction and mixed learning mechanism, the agent decision-making flexibility is enhanced; dynamic data are efficiently processed and deep association is mined by utilizing incremental topology analysis and a knowledge graph; and through anti-disturbance and propagation simulation, a chaos edge emergence phenomenon is captured, and finally, deduction and actual combat deviation is reduced through multi-module cooperation.

The invention provides a network topology relation graph dynamic generation method, which belongs to the technical field of network topology, and comprises the following steps: starting an improved Tarjan strong connectivity component algorithm in each subnet partition to carry out loop detection, executing distributed loop verification on cross-subnet boundary nodes, collecting ARP table entries, routing table entries, OpenFlow flow tables and virtualization platform interface data, and sending the ARP table entries, the routing table entries, the OpenFlow flow tables and the virtualization platform interface data to each subnet partition; the method comprises the following steps: performing power spectrum analysis on an optical power monitoring point on optical network equipment to identify a wavelength channel occupancy state, reconstructing a wavelength routing path of a DWDM system, and performing two-stage submission on topological data acquired by distributed detection nodes by adopting a causal consistency protocol to synchronously generate a global unified network topological relation graph; the technical problems that the incremental updating efficiency is low and the cross-layer mapping relation is difficult to reconstruct in real time when the network topology dynamically changes are solved.

The invention provides a VPN networking system and method, electronic equipment and a computer program product. Relates to the technical field of communication. In view of the problems of traditional VPN link resource waste and difficult session hot backup, the invention provides a VPN networking system, which comprises a VPN site and a VPN gateway cluster, and the VPN gateway cluster comprises M device nodes based on mesh distribution; a VPN tunnel is established between each device node and the VPN site; the VPN subnet at the local area network side is divided into M subnet segments, and each equipment node bears one subnet segment; every two device nodes in the VPN gateway cluster are in communication connection based on a VPN link, each device node comprises routing information of M device nodes, the routing information is used for indicating a forwarding path of sub-network segment flow when a system breaks down, and the routing information comprises next hop information of a sub-network segment borne by each device node. According to the embodiment of the invention, the method does not depend on session hot backup, achieves the quick and smooth switching of the VPN tunnel, and improves the resource utilization rate.

A wireless communication method includes: obtaining subnet information of a first subnet corresponding to a first session; receiving a data packet, where the data packet is a multicast packet or a broadcast packet; determining, based on the subnet information, that the data packet corresponds to the first session in the first subnet; and sending the data packet via the first session.

The invention relates to the technical field of video analysis and processing, and discloses a key frame extraction method based on hierarchical attention and potential gating and related equipment. The method comprises the following steps: analyzing an original video stream and separating frames, adjusting the resolution and normalizing to obtain a standardized sequence; inputting the standardized sequence to a multi-stage convolutional neural network encoder, generating multi-stage spatial features, and obtaining a spatial attention map through attention mapping; pooling the lowest resolution feature to obtain a frame-level embedded vector sequence, and performing time sequence conversion to obtain time sequence attention association information; constructing a potential gating sub-network containing a strategy network, wherein the strategy network outputs spatial feature weights of all levels; the fusion unit sums the spatial features and the corresponding attention maps according to weights, and fuses time sequence results to obtain a fusion feature sequence; and outputting a frame importance score after the scoring network processing, and screening according to the importance score to obtain a target key frame. According to the method, key frame extraction accuracy is improved, and redundant frames are effectively reduced.

The invention relates to the technical field of data processing, and discloses a multi-modal data processing method, device, equipment and medium, which can be applied to the financial and medical fields, the method comprises the following steps: preprocessing multi-modal input data to obtain standardized multi-modal data, and constructing a training data set for simulating missing modalities; carrying out feature generation on existing missing modalities to construct a complete modal feature set; and inputting each modal feature in the complete modal feature set into the corresponding independent processing sub-network, and fusing to obtain a multi-modal data processing result. In the invention, aiming at the problem that available modal information cannot be effectively utilized to supplement a missing modal in a complex scene in which multi-modal data are simultaneously missing in the prior art, a generative completion model can be utilized to perform feature generation on the existing missing modal, and the feature is input to a corresponding independent processing sub-network; and finally, outputting a multi-modal data processing result, so that the decision accuracy of the model in a modal missing scene is improved.

The invention discloses a reverse knowledge distillation-based cross-stage feature fusion method for industrial anomaly detection and positioning, and relates to the technical field of knowledge distillation and transfer learning. The method comprises the following steps: step 1, constructing a teacher network encoder, a student network encoder and a decoder; 2, extracting input image features through a student network encoder, and reconstructing the features through a student decoder; step 3, student network cross-stage feature fusion design; 4, locally sensing a dynamic attention module LDA, and adding a sliding window and convolution to replace original global pooling; and step 5, a self-adaptive multi-scale feature fusion module CAAMS-FF establishes a dependency relationship of a context so as to realize a fine-grained high-quality feature reconstruction effect. And step 6, inputting an anomaly graph obtained by the teacher-student network into the segmentation sub-network to obtain a final anomaly detection and positioning result. While the calculation efficiency is maintained, the spatial positioning precision and the multi-scale adaptability are significantly improved.