101 results about "Tree network" patented technology

The invention provides a multi-level attribution and recommendation method and system for medical management decisions, and relates to the technical field of medical management decisions, and the method comprises the steps: extracting graph node evolution characteristics through a multi-level medical data knowledge graph and a decision knowledge base by employing a time sequence diagram convolutional network, constructing a decision influence propagation graph through a cross-layer interactive attention network, and carrying out the analysis of the decision influence propagation graph; calculating the contribution degree of each decision node to the medical effect, generating an attribution analysis result, determining a decision influence hierarchical system according to the attribution result, and generating and evaluating candidate decision paths in combination with a multi-objective optimization network, a hierarchical reinforcement learning network and a Monte Carlo search tree network. In conclusion, accurate attribution analysis and multi-objective optimization of the management decision can be realized, the hospital resource allocation efficiency is effectively improved, the operation cost is reduced, scientific decision support is provided for a hospital management layer, and the hospital fine management level and the overall operation efficiency are remarkably enhanced.

The invention provides a petrochemical pipeline full-life-cycle detection and evaluation method and equipment. The method comprises the following steps: constructing a pipeline detection digital twin model comprising a geometric model and material attributes; collecting detection data of the pipeline through the deployed multi-source sensor network; inputting the detection data into the pipeline detection digital twin model to perform feature extraction and analysis of the pipeline state; constructing a heterogeneous tree network and executing hierarchical evaluation; optimizing the maintenance plan based on the evaluation result; and predicting the residual life of the pipeline. Through the digital twinning technology and multi-source detection data fusion, full-life-cycle accurate evaluation of the petrochemical pipeline is realized.

The embodiment of the invention provides a document-based intelligent question and answer method and device, electronic equipment and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: performing document analysis on a target electronic document to obtain document attributes and document contents; based on the document attribute and the document content, document segmentation is conducted on the target electronic document, document segmentation units and unit position information of the document segmentation units are obtained, and the document segmentation units comprise segmentation unit identifiers; based on the document attribute, performing network construction on the target electronic document to obtain a document structure tree network; on the basis of the unit position information, the document segmentation units are arranged in sequence, and a document linear list is obtained; performing document query on the document structure tree network or the document linear list based on the document problem information and the segmentation unit identifier to obtain document query information; and based on the document query information, performing question reply on the document question information. According to the embodiment of the invention, the intelligent question and answer accuracy based on the document can be improved.

Provided in the present application is a data routing system. The system comprises a first topological network, a second topological network and a third topological network, wherein the network structure of the first topological network is a toroidal network topology structure consisting of a first node group, the first node group comprises a first number of network nodes, and the network nodes in the toroidal network topology structure have connection relationships with upstream nodes and downstream nodes in each dimension; the second topological network is a fat-tree network topology structure consisting of the network nodes in the first topological network and a switch; there is a cascade relationship between the network nodes in the first topological network and the switch; the first topological network is nested in the third topological network, and the network structure of the third topological network is a fully connected network topology structure consisting of a second node group; and the second node group comprises a second number of first node groups, and there is a connection relationship between the first node groups. Thus, large-scale or ultra-large-scale networking can be realized at a relatively low cost.

The invention provides an insurance product pushing method and system for modeling by using user interests, which can accurately capture relevance and characteristics between user behaviors by constructing target interaction behavior graph structure data, and perform user interest modeling by using a logistic regression network and a random forest tree network respectively, so as to improve the user interests. And two types of modeling data are fused, so that the interests, preferences and potential demands of the user can be reflected more comprehensively. In addition, network parameter learning is carried out by adopting different example product interaction behavior data sequences, so that the accuracy and robustness of user interest modeling are further improved. Finally, insurance product pushing is carried out based on the target user interest modeling data, the pushing accuracy and the user satisfaction degree can be remarkably improved, and data support is provided for personalized and intelligent recommendation of insurance products.

The invention provides an Ethernet network card fault prediction and recovery method and related equipment, and the method comprises the steps: carrying out the feature fusion of a network card physical layer parameter, a protocol stack flow parameter and environment interference data of an Ethernet, and generating a joint feature vector; determining a quantile prediction matrix corresponding to the network card key performance index according to a dynamic quantile prediction model constructed based on the joint feature vector; analyzing a node fault probability according to the quantile prediction matrix, and determining a high-risk node list based on a preset dynamic entropy threshold; constructing a dynamic tree network model through the high-risk node list, and generating a candidate recovery scheme set based on physical switching and protocol reconstruction; and optimizing the candidate recovery scheme set through an improved genetic algorithm, and outputting a global optimal recovery scheme. Through the implementation of the scheme of the invention, the potential risk can be predicted before the fault occurs, the optimal recovery path can be quickly matched, and the stability of the network and the continuity of the service are ensured.

The present invention provides a wireless communication method and system for the Internet of Things based on the Star Flash protocol. The method comprises the following steps: deploying a plurality of wireless communication nodes in a wide area or a long-distance path to ensure that the distribution distance between any two of the wireless communication nodes does not exceed 200 meters in the absence of obstacles; selecting the wireless communication nodes as gateway nodes, which have the ability to communicate with any other wireless communication nodes; assigning a unique identification code to each of the wireless communication nodes, grouping them by the identification code, so that the wireless communication nodes with the same identification code can automatically construct a tree network structure under the guidance of the designated gateway node; using the wireless communication nodes to connect to the opposite end according to the signal strength and the current network topology structure, and after the connection is established, using the MAC address to access and control data, thereby reducing costs and improving communication reliability and stability.

The invention relates to a clock synchronization system and method based on dual-mode communication, the system is a tree network and comprises a network main node CCO, a proxy node PCO and a child node STA device, the dual-mode communication is high-speed carrier and high-speed wireless, a time synchronization sub-layer is added in a data link layer, and the time synchronization sub-layer is used for synchronizing the data link layer. The frame structure of the time synchronization sub-layer comprises a time synchronization level for representing a clock synchronization priority; time synchronization operation of the system is carried out in a high-speed carrier channel and a high-speed wireless channel at the same time, transmitted time synchronization signals comprise time synchronization priority signals containing time synchronization levels, and synchronization of PCO and STA clocks and CCO clocks in the same low-voltage transformer area dual-mode communication network is achieved based on the time synchronization priority signals. Compared with the prior art, the method has the advantages of improving the clock synchronization efficiency, precision and robustness and the like.

The invention relates to the technical field of forest resource investigation and point cloud processing, and discloses a laser radar individual tree segmentation method based on point cloud dimensionality reduction and individual tree network reforming, and the method comprises the steps: taking airborne LiDAR point cloud as a special network structure, replacing a single point with a point cloud cluster, and taking a fitting cluster circle as a hub. Dimensionality reduction processing from the three-dimensional point cloud to the two-dimensional hub is carried out, and redundant information is reduced. Three hub clustering methods (Euclidean distance, four-dimensional similarity and cosine similarity) are provided to construct individual tree hub clusters, and clustering results are mapped back to the three-dimensional point cloud to realize individual tree segmentation. The calculation pressure caused by huge point clouds is effectively relieved, and meanwhile, the segmentation precision and universality are improved; through a multi-strategy hub clustering method, the adaptability and robustness in a complex forest environment are improved; by introducing automatic precision evaluation based on convex hulls, low efficiency and subjectivity of traditional manual discrimination are avoided, objectification and automation of segmentation precision are realized, and the method has high theoretical value and application prospect.

A fault-tolerant quantum computer using topological codes such as surface codes can have an architecture that reduces the amount of idle volume generated. The architecture can include qubit modules that generate surface code patches for different qubits and a network of interconnections between different qubit modules. The interconnections can include “port” connections that selectably enable coupling of boundaries of surface code patches generated in different qubit modules and / or “quickswap” connections that selectably enable transferring the state of a surface code patch from one qubit module to another. Port and / or quickswap connections can be made between a subset of qubit modules. For instance port connections can connect a given qubit module to other qubit modules within a fixed range. Quickswap connections can provide a log-tree network of direct connections between qubit modules.

A network on chip for high performance computing includes a first fat-tree network topology that connects each of a number of initiator nodes to one of a number of initiator core switches through one or more levels of initiator edge switches. A second fat-tree network topology connects each of the target nodes to one of a number of target core switches through one or more levels of target edge switches, and the first fat-tree network topology is joined to the second fat-tree network topology through connections between the initiator core switches and the target core switches.

The invention relates to the technical field of satellite-ground laser communication, and discloses a banyan-shaped network communication method and system based on multi-beam multipath transmission and space diversity synthesis aperture reception. The method comprises the following steps: dividing a light beam into a plurality of beams at a satellite-based terminal through beam forming to point to a plurality of ground stations; multiple ground stations receive data in parallel and then gather the data to a data center for time-space synchronization; and the data center adopts a multi-branch neural network to intelligently fuse the synchronized data so as to reconstruct a signal. According to the invention, by constructing the satellite-to-ground collaborative architecture, signal interruption and data loss caused by atmospheric turbulence are effectively suppressed, the limitation of high-error-code communication is broken through, and remote, high-speed and high-reliability satellite-to-ground communication is realized.

The invention discloses an autonomous construction method and system of a spectrum prediction analysis model, which can effectively improve the generalization performance of the model by adopting a pruning strategy to remove redundant model branches while autonomously selecting modeling variables and exploring different model structures. According to the technical scheme, an analysis data set is constructed and randomly divided into a training set and a test set; establishing a partial least square model by using the training set data, selecting initial variables according to the importance of regression coefficients, and adaptively adjusting the maximum tree depth according to the number of the initial variables; randomly generating an initial population according to the function set and the terminal set of the tree network; calculating the error reduction rate of tree network sub-trees, pruning, and calculating the fitness of pruned tree individuals; performing selection, crossover and mutation operation on individuals according to the individual fitness of the population and preset parameters, performing individual fitness evaluation again and updating the optimal individual; and forming a spectrum prediction analysis model by using the obtained optimal individual, and predicting a test set sample.

A fault-tolerant quantum computer using topological codes such as surface codes can have an architecture that reduces the amount of idle volume generated. The architecture can include qubit modules that generate surface code patches for different qubits and a network of interconnections between different qubit modules. The interconnections can include “port” connections that selectably enable coupling of boundaries of surface code patches generated in different qubit modules and / or “quickswap” connections that selectably enable transferring the state of a surface code patch from one qubit module to another. Port and / or quickswap connections can be made between a subset of qubit modules. For instance port connections can connect a given qubit module to other qubit modules within a fixed range. Quickswap connections can provide a log-tree network of direct connections between qubit modules.

It realizes real-time perception of forest ecosystems, reduces the cost and risk of resource surveys, and improves inspection accuracy and quality. [Solution] The forest tree network system includes a forest tree network design module, a forest IoT monitoring station construction module, a forest tree network operation and collection module, a forest resource assessment module, a forest tree network monitoring early warning module and a forest tree network scheduling and maintenance module. Through the design and construction of the station, high-frequency monitoring of the growth status of forest trees in the ecosystem is performed, and important indicator data such as forest growth volume, consumption volume, withering volume, forest accumulation volume per unit area and forest biomass are directly output, so as to carry out the adjustment, scheduling, control and protection of the forest ecosystem, improve the accuracy and quality of resource surveying, reduce monitoring costs and risks, and enhance the restoration accuracy and timeliness of forest ecosystem protection.

A deep well ventilation system is a core link for guaranteeing safe production of a mine and is also a key technical challenge for realizing efficient and green mining of an ultra-large deep well mine. Ventilation energy-saving optimization is an important subject which must be solved by deep well mining. In a deep well ventilation system, reasonable configuration of a communication network structure is a basic guarantee for achieving real-time monitoring of underground environment data and reliable transmission of control instructions. The main contradiction faced by the current deep well communication network design lies in that the network redundancy design introduced for improving the reliability of data transmission can enhance the fault-tolerant capability of the system, but also causes the problems of complex network topology structure and increased construction cost at the same time. In order to solve the technical problem, the invention provides a communication network design and configuration method suitable for a mine ventilation scene. According to the method, a quantitative relation model of the network construction cost and the safety performance is innovatively established, and the optimal network node configuration scheme is determined through target optimization. Particularly, according to the method, the high reliability of the ring network and the high efficiency of the tree network are fused, so that the optimal design of the communication network topology structure is realized, and the network economy is remarkably improved while the system reliability is ensured.

The invention relates to the technical field of coal mine underground wireless communication, in particular to a time slot distribution and dormancy control method and system of a low-power-consumption wireless sensor network. The invention discloses a time slot allocation and dormancy control method for a low-power-consumption wireless sensor network, and the method comprises the following steps: S1, constructing a multi-hop tree network topology which takes a gateway as a center, and determining the neighborhood relation and the optimal path of each sensing node; s2, defining system parameters; s3, distributing native time slot numbers for all the sensing nodes; s4, distributing relay time slot numbers for all non-tail-end nodes, ignoring the tail-end nodes during distribution, and adopting a distribution principle which is the same as the original time slot number in the step S3; s5, all the sensing nodes work according to rules; and S6, the common dormancy time Tc = Tp-Nt * Ts is calculated, and Nt is the total time slot number required for completing data transmission of the whole network. Asynchronous time slot allocation and dynamic dormancy control are adopted, and the problems of high energy consumption, large interference and long delay of a synchronization scheme are solved through time-sharing awakening, conflict isolation and relay optimization.

The invention discloses a substrate defect detection method based on an incremental learning CNN tree model, and the method comprises the following steps: constructing a source node network, enabling multiple layers of nodes to form an initial node network in a parallel cascading manner, and enabling each node to be provided with a substrate defect classification CNN network; calculating a three-dimensional likelihood matrix of the input substrate image, performing descending sorting on value elements of the three-dimensional likelihood matrix, and establishing a corresponding Softmax likelihood value table for each sub-node; adding a defect class to be detected to sub-nodes of the initial node network by using a semi-supervised growth rule, and adapting to a new substrate defect sample class or a new-old combined substrate defect sample class in a tree-shaped growth mode; retraining or finely adjusting a hidden layer of the substrate defect classification CNN network based on a back propagation strategy; and based on the substrate defect classification CNN network of each node of the trained tree network, outputting a substrate defect type result. According to the invention, accurate and rapid substrate defect detection can be realized, and the robustness is better.

Provided in the present application is a data routing system. The system comprises a first topological network, a second topological network and a third topological network, wherein the network structure of the first topological network is a toroidal network topology structure consisting of a first node group, the first node group comprises a first number of network nodes, and the network nodes in the toroidal network topology structure have connection relationships with upstream nodes and downstream nodes in each dimension; the second topological network is a fat-tree network topology structure consisting of the network nodes in the first topological network and a switch; there is a cascade relationship between the network nodes in the first topological network and the switch; the first topological network is nested in the third topological network, and the network structure of the third topological network is a fully connected network topology structure consisting of a second node group; and the second node group comprises a second number of first node groups, and there is a connection relationship between the first node groups. Thus, large-scale or ultra-large-scale networking can be realized at a relatively low cost.

The embodiment of the invention provides a network-on-chip, a bare chip, a chip and a routing method. The network-on-chip comprises a plurality of first routing nodes, a plurality of second routing nodes and a plurality of third routing nodes, the plurality of first routing nodes are connected to form a grid network with multiple rows and two columns, each row of the grid network comprises two first routing nodes, and each column of the grid network comprises a plurality of first routing nodes; the first routing nodes located in the first column of the mesh network are connected with the second routing nodes, and different first routing nodes are connected with different second routing nodes; the first routing nodes located in the second column of the mesh network are connected with the third routing nodes, and different first routing nodes are connected with different third routing nodes; at least part of the second routing nodes are connected with at least one main processing node through a tree network; at least part of the third routing nodes are connected with at least one slave processing node through the tree network. According to the scheme, the network-on-chip can meet the requirement of high bandwidth and occupies a small chip area.

Embodiments of the disclosure provide a multi-hop wireless network system including multiple network nodes, a method for maintaining a tree network topology, and a first member node. The multi-hop wireless network system includes the first member node. The first member node is connected to a control node of the network nodes through a first relay node in the tree topology network formed by the network nodes, and operates as a second relay node. The first member node is configured to: in response to determining that connection between the first relay node and the first member node is disconnected, stop operating as the second relay node.

The application discloses a kind of support multi-mode convolution's memory computing integrated acceleration circuit, chip and equipment, the memory computing integrated acceleration circuit of the application includes input buffer (1), column address decoder (2) and 64 multiply-accumulate operation arrays (3), the multiply-accumulate operation array (3) includes: row address decoder (31), multi-mode adaptability controller (32), SRAM array (33), multiplier (34) and addition tree network (35), the 64 multiply-accumulate operation arrays (3) share one common column address decoder (2) and are connected to same input buffer (1).The application aims at using the flexibility and scalability inherent in digital CIM, without sacrificing the original CIM storage density, to design efficient control logic, to support various convolutions with minimal additional area cost and maintain high computing efficiency.

Provided in the present application is a data routing system. The system comprises a first topological network, a second topological network and a third topological network, wherein the network structure of the first topological network is a toroidal network topology structure consisting of a first node group, the first node group comprises a first number of network nodes, and the network nodes in the toroidal network topology structure have connection relationships with upstream nodes and downstream nodes in each dimension; the second topological network is a fat-tree network topology structure consisting of the network nodes in the first topological network and a switch; there is a cascade relationship between the network nodes in the first topological network and the switch; the first topological network is nested in the third topological network, and the network structure of the third topological network is a fully connected network topology structure consisting of a second node group; and the second node group comprises a second number of first node groups, and there is a connection relationship between the first node groups. Thus, large-scale or ultra-large-scale networking can be realized at a relatively low cost.

A Mesh network-based synchronization system includes a plurality of stage lights and a controller for controlling the stage lights, each of the stage lights having an ad hoc network module. A LAN is established based on the Mesh technology among several stage lights by the controller under the help of the ad hoc network module, one stage light in the LAN acts as a master node, the remaining stage lights forms at least one tree-shaped network topological structure connected to the master node, and some stage lights are selected to form a group. Subsequent to time synchronization of the stage lights in the group, delay time of each stage light relative to the master node is respectively calculated according to the layer level, the master node sends a start command in advance of the corresponding delay time to the respective stage light to synchronously display a show.

An apparatus comprises: an input port configured to receive an optical wave; a plurality of optical amplification modules (OAMs), each configured to apply a gain to an optical wave; a plurality of optical power splitting modules (OPSMs), each configured to provide portions of an optical wave to two or more outputs according to an optical power splitting ratio; and a plurality of optical ports, each configured to provide an optical wave; wherein the plurality of OPSMs is interconnected in a tree network that comprises: a root stage comprising a path segment between the input port and a first OPSM, a final stage comprising a plurality of path segments between OPSMs and optical ports, and a plurality of inner stages, each comprising a plurality of path segments between an upstream and a downstream OPSM; and wherein at least two inner stages each include an OAM on one or more path segments.

The application discloses a power distribution line single-phase tree contact grounding fault risk assessment method and system, and the method comprises the following steps: a fractal tree network model is used to simplify and simulate the real shape of a tree to build a tree model; the initial moisture content of the tree model is obtained, and the temperature and transition resistance of the tree model are continuously monitored; a single-phase tree contact fault model is built, a conductor phase voltage is obtained, the current heat production, heat dissipation and heat absorption of the tree are continuously calculated, the water evaporation quality is calculated and obtained, and the moisture content of the tree is updated using the water evaporation quality; the binary relationship between the conductivity, moisture content and temperature of the tree model is fitted and obtained; the binary relationship is applied to the iterative calculation of the fault model to obtain the highest temperature of the tree in the whole process; and the risk level of the tree is assessed according to the highest temperature by taking a set temperature threshold value. The application can accurately realize the fire risk assessment of the size of the tree, and is helpful to the safe operation of the power grid and the protection of the ecological environment.

The application discloses a memory-based XNOR operation data processing method, and belongs to the technical field of computer data processing. The method comprises the following steps: obtaining two or more binary input data to be subjected to XNOR operation and preprocessing to obtain bit-level addressing data; configuring a mapping storage array pre-storing XNOR operation results according to the bit-level addressing data; generating parallel addressing signals; triggering multiple storage units in the mapping storage array to simultaneously output parallel result data based on the parallel addressing signals; and receiving the parallel result data and aggregating to obtain a final XNOR operation result. The application adopts a multi-dimensional storage unit array integrated micro logic gate to realize local XNOR operation, combines an innovative address coding scheme to optimize addressing energy consumption and data reliability, and through an efficient parallel addressing signal generation and multi-level tree network aggregation mechanism, can embed XNOR operation logic into the storage, and significantly improve the parallelism, energy efficiency ratio and operation speed of data processing.

A network on chip for high performance computing includes a first fat-tree network topology that connects each of a number of initiator nodes to one of a number of initiator core switches through one or more levels of initiator edge switches. A second fat-tree network topology connects each of the target nodes to one of a number of target core switches through one or more levels of target edge switches, and the first fat-tree network topology is joined to the second fat-tree network topology through connections between the initiator core switches and the target core switches.

A bandwidth estimation system includes an interface and a processor. The interface communicates with a fat-tree network (FT network) containing multiple switches, including (i) multiple leaf switches belonging to a lower level, (ii) multiple spine switches belonging to an upper level, and (iii) multiple intermediate switches belonging to one or more intermediate levels between the upper and lower levels. Links connect selected switches.The processor is to calculate, for a given level of the FT network which is divided into several groups of switches, one or more oversubscription ratios for each of the groups, where an oversubscription ratio of a group represents a ratio between (i) a first available bandwidth on the links connecting the switches in the group to a lower level, and (ii) a second available bandwidth on the links connecting the switches in the group to a higher level, and report a quality factor of the FT network based on the oversubscription ratios.

The utility model discloses a place lighting control system which comprises a first system and a second system. The first system comprises a tree network and a control platform, the control platform is connected with a root node of the tree network, a part of leaf nodes of the tree network are a first brightness sensing device, a human body sensing device and a first actuator for controlling on and off of lamps, and the first actuator is arranged in a people flow non-dense area; and the second system is arranged in a crowded area and comprises a communication control module, a second brightness sensing device and a second actuator for controlling on and off of the lamp, and the second actuator and the second brightness sensing device are both connected with the communication control module. According to the utility model, the first system and the second system arranged in the crowded area are respectively arranged according to the crowded non-dense degree of the place area, so that the flexible deployment of the system is realized, the lighting efficiency is improved, and the first system adopts a tree network, so that the system can be conveniently expanded or reduced according to the change of the place condition.