70 results about "Network characterization" patented technology

Deployment of an application is managed using available resources from a resource system. A resource dependency characterization of the application is established to enable deployment of the application. A logical characterization and a network characterization of a desired deployment are also established. A deployment plan containing an outline for resources and configurations used for deployment is developed based on the resource dependency characterization, the logical characterization and the network characterization. The deployment plan is processed for correlation with the available resources. The processed deployment plan is then implemented.

A method of representing and using fractures in a model of a subterranean reservoir is described including the partitioning the fracture network into a discretely modeled part and a remaining statistically described part from a statistical description of all fractures, the determination of the correlation effects caused by fractures with dimensions exceeding dimension of the local grid cells and the determination of petrophysical properties while allowing for arbitrary distribution of fracture orientations, with all three aspects being combinable to improve the modeling of fractures and the simulation of fractured reservoirs.

Deployment of an application is managed using available resources from a resource system. A resource dependency characterization of the application is established to enable deployment of the application. A logical characterization and a network characterization of a desired deployment are also established. A deployment plan containing an outline for resources and configurations used for deployment is developed based on the resource dependency characterization, the logical characterization and the network characterization. The deployment plan is processed for correlation with the available resources. The processed deployment plan is then implemented.

A method of selecting a gateway node in a remote network and for handing over to the selected gateway node. Nodes in a remote energy aware network connect through a gateway node to a backbone network in an interplanetary communications network. Each node optimizes a stability function describing communications to neighboring nodes and to the backbone. Optimization is for maximum network stability and for efficient node energy consumption. Optimization identifies a handover time and nodes initiate handover sufficiently in advance of the identified handover time to complete at that time. Nodes continually monitor and update network characterization parameters to identify a next optimal handover time.

A method of characterizing a wiring network is implemented in a system which includes a test controller and at least two probes. On commands from the test controller, at least one of the probes changes its impedance between the nominal impedance of the wiring network and a mismatch impedance. Reflectometry measurements are performed before and after of switching the impedance of the second probe. At the first probe, an RF signal is generated and a reflected signal is measured. Then, the impedance of the second probe is changed, and again an RF signal is generated and a reflected signal is measured at the first probe. Additionally, a frequency response may be measured at the second probe. The results of the measurements are used for characterization of a transmission line between the first and second probes

The invention provides a network representation learning method based on a meta-structure and a graph neural network. According to the method, the information of the neighbor nodes is aggregated through the graph neural network and the weighted attention mechanism, the candidate meta-structure set is generated through the hierarchical search algorithm, meta-structures do not need to be defined inadvance, and compared with a previous meta-path, the method can consider more complex structure information between the nodes. According to the method, the strong learning ability of the graph neuralnetwork and the rich semantics of the meta-structure are fused, and the problems that an existing meta-path-based method is single in consideration structure and needs to depend on experience to specify the meta-structure are effectively solved. Moreover, the introduction of a weighted attention mechanism can explicitly consider the quantity information in the meta-structure. And a final node which is more accurate than a result of a traditional representation classification mode is generated, and the final node can be used as vector representation to be used in subsequent other machine learning figures.

The invention discloses a rapid network representation learning algorithm based on a width learning system. The method comprises the following steps of S1, importing a network graph module based on atext, parsing and storing a network topological structure in a dictionary format, wherein keys in the dictionary represent network nodes, values corresponding to the keys form a list and represent a node sequence at the other end of the edge where the nodes are located; S2, performing random walk on the network nodes to generate a walk sequence; S3, constructing a network representation learning model based on a width learning system, taking the walking sequence generated in the step S2 and a representation vector with the dimension of K as input, generating a feature vector of a network nodein a feature vector layer, and enhancing the nonlinear classification capability of a network representation learning model by introducing an activation function in an enhancement vector layer to finally realize text-based network multi-label classification. A width learning system model is adopted in the algorithm, and representation learning of network nodes can be rapidly completed.

The invention discloses a homonymous author disambiguation method based on network representation and semantic representation. The method comprises the following steps: 1) extracting semantic featuresand discrete features of each paper in a target paper library; 2) calculating the similarity among the papers based on the discrete features to obtain a relationship similarity matrix of the papers;if one paper and other papers do not have a co-author or institution, adding the paper and other papers into an outlier paper set; 3) calculating a semantic similarity matrix of the papers based on the semantic features of the papers; adding papers without semantic features in the target paper library into an outlier paper set; 4) performing weighted summation on the relationship similarity matrixand the semantic similarity matrix to obtain a paper similarity matrix, and clustering the paper similarity matrix; adding papers which do not belong to any cluster into the outlier paper set; and 5)distributing the papers in the outlier paper set to the corresponding clusters by utilizing a similarity threshold matching-based method. According to the method, high-accuracy disambiguation of authors with the same name of the papers is realized.

Some embodiments include a method for determining whether to establish a connection between a computer and a network. In some embodiments, the method includes: receiving first network information from a network; determining a first trust level for the network by matching the network information to provisioning information, wherein the provisioning information indicates the first trust level and an authentication method of the network; verifying that network supports the authentication method, wherein the verifying includes exchanging information with the network; after verifying that the network supports the authentication method, establishing a connection to the network, wherein the connection provides access to an Internet.

The invention discloses an online transaction multi-subject behavior modeling method based on heterogeneous network representation learning. The modeling method comprises the following steps: step 1,expressing original online transaction data as a heterogeneous network by utilizing an association map, and carrying out data supplement on missing information in the heterogeneous networked transaction data by utilizing heterogeneous network representation learning to establish an individual behavior model; And step 2, establishing behavior models of multiple subjects with different dimensions byutilizing multi-subject behavior modeling, integrating the multiple subject behavior models to obtain a judgment result, and predicting the possibility of transaction abnormity. The method can be used for detecting the fraudulent transaction of the online transaction, intercepting the fraudulent transaction and protecting the fund safety of a user and an enterprise.

A method of characterizing a wiring network is implemented in a system which includes a test controller and at least two probes. On commands from the test controller, at least one of the probes changes its impedance between the nominal impedance of the wiring network and a mismatch impedance. Reflectometry measurements are performed before and after of switching the impedance of the second probe. At the first probe, an RF signal is generated and a reflected signal is measured. Then, the impedance of the second probe is changed, and again an RF signal is generated and a reflected signal is measured at the first probe. Additionally, a frequency response may be measured at the second probe. The results of the measurements are used for characterization of a transmission line between the first and second probes.

The invention discloses a personalized commodity recommendation method and system based on multilayer heterogeneous attribute network representation learning, and the method comprises the steps: taking an interaction behavior between a user and a commodity as an edge, constructing a multilayer heterogeneous attribute network, and carrying out the decoupling of the multilayer heterogeneous attribute network into a plurality of simple binary networks; performing weighted accumulation on the adjacency matrixes of all the binary networks to obtain a newly combined adjacency matrix, and performingspectrogram conversion; fusing the adjacency matrix and the node attribute characteristic matrix after spectrum conversion, and finally obtaining representation vectors of all nodes by using a randomprojection method; obtaining a verification set from the historical data to perform parameter adjustment, and obtaining a representation vector of each node; and measuring the preference of the user to the commodity by utilizing cosine similarity so as to carry out personalized recommendation. According to the invention, various interaction behaviors between the user and the commodity are considered at the same time; the interactive relationship among various behaviors can be captured without human intervention; attribute information of users and commodities is effectively fused; network representation learning is carried out by using random projection so that the method efficiency is greatly improved and the recommendation performance is improved.

A medical image super-resolution reconstruction method based on multi-attention residual feature fusion combines two modules of a channel attention mechanism and a space attention mechanism, and enables a network to perform weighting processing on features of an input image, thereby solving the problem of insufficient network characterization capability caused by equal processing of the input features. A residual attention feature fusion module (RAFF) is provided, local features in the neural network are fused, and global features are fused, so that local residual features in the whole network can be more fully utilized, and the problem that the deep neural network is too deep and the learning ability of the network is insufficient is solved. According to the method, super-resolution reconstruction is effectively carried out on the medical image, and the image super-resolution reconstruction performance of a deep super-resolution network can be improved.

Embodiments of a method of calculating the equivalent series resistance of a matching network using variable impedance analysis and matching networks analyzed using the same are provided herein. In one embodiment, a method of calculating the equivalent series resistance of a matching network includes the steps of connecting the matching network to a load; measuring an output of the matching network over a range of load impedances; and calculating the equivalent series resistance of the matching network based upon a relationship between the measured output and the load resistance. The load may be a surrogate load or may be a plasma formed in a process chamber.

A method of representing and using fractures in a model of a subterranean reservoir is described including the partitioning the fracture network into a discretely modeled part and a remaining statistically described part from a statistical description of all fractures, the determination of the correlation effects caused by fractures with dimensions exceeding dimension of the local grid cells and the determination of petrophysical properties while allowing for arbitrary distribution of fracture orientations, with all three aspects being combinable to improve the modeling of fractures and the simulation of fractured reservoirs.

The invention relates to the technical field of network characterization, in particular to a network characterization method based on cross-double-layer network random walk, which comprises the following steps of: A) establishing a network topological structure; B) obtaining a role similarity matrix; C) establishing a double-layer hybrid network; D) obtaining a random walk sequence; And E) obtaining the representation of the network by using the continuous bag-of-words model. The method for establishing the role similarity matrix S comprises the following steps of: B1) listing all subgraphs with sizes smaller than or equal to a given size k; B2) enumerating a non-isomorphic track, and recording the number as m; B3) expressing the condition that each node participates in m roles by using avector with the length of m; And B4) taking the similarity of the role representation vectors of every two nodes as the similarity of the two nodes and a role similarity matrix S. The method has the advantages that by means of the random walk model and the continuous bag-of-words model, representation of network adjacency and structural similarity is fused at the same time, and effective representation of non-connected network nodes with similar roles can be achieved.

The invention provides a fungus image recognition method and device, electronic equipment and a storage medium. The method comprises the steps that at least one base network of a pre-trained fusion network model is used for extracting features of a plurality of fungus images, fusion features are obtained, the base network represents a neural network for conducting convolution operation and fusionoperation on the images, and the angles or growth periods of fungi in the fungus images are different; and the classification layer of the fusion network model is used to identify the fusion featuresto obtain the classification of the fungus images. In the implementation process, the features of the fungus images are extracted by using at least one base network of the fusion network model to obtain the fusion features, and the classification layer of the fusion network model is used for identifying the fusion features to obtain the classification of the fungus images, so that the efficiency of identifying the fungus images is effectively improved.

The invention discloses a grain depot personnel nonstandard operation detection method based on an improved YOLOv3 algorithm. The grain depot personnel nonstandard operation detection method comprises the following steps: carrying out image acquisition on grain depot personnel nonstandard operation behaviors; inputting the acquired image into a grain depot non-standard operation behavior detection network in an upper computer; outputting a result picture with an accurate mark of the non-standard operation behavior of the personnel; and displaying the result and storing same in the upper computer. The grain depot non-standard operation behavior detection network comprises a trunk layer constructed based on a YOLOv3 network and a feature fusion output layer, and a scale context selection attention module SCA is embedded in a Y2 layer and a Y3 layer of the feature fusion output layer respectively. The invention overcomes the defects in the prior art, and provides the improved YOLOv3 algorithm-based detection method for the irregular operation of the grain depot personnel, which is more suitable for grain depot scene detection and has stronger network characterization capability.

The invention discloses a network service anomaly detection method and device based on attribute network representation learning, and the method comprises the steps: obtaining initial network servicedata, constructing a heterogeneous information network based on the initial network service data, and obtaining a node attribute set; constructing an attribute vector set based on the node attribute set, and constructing an attribute information network according to the attribute vector set and the heterogeneous information network; constructing a target function based on the attribute informationnetwork, and constructing a mapping relationship between nodes in the attribute information network and vector representations corresponding to the nodes based on vectors corresponding to the nodes to be learned in network representation learning obtained by solving the target function; and training based on the training set data to obtain an anomaly detection model, and calculating an anomaly probability of each piece of network service data in the test set data according to the anomaly detection model. According to the method, the relevance of the nodes in the attribute information networkis enhanced, the generalization ability of the anomaly detection model is improved, and better guarantee is provided for anomaly detection, anomaly interception and fund security protection of users and enterprises.

The invention discloses a multi-layer fusion beacon detection and path restoration method and device. The method comprises the following steps: implanting a multi-layer fusion beacon in a target system network routing node data packet; obtaining a routing node network data packet of an implanted beacon in a target system; network characterization learning is carried out on the network data packet with the implanted beacons, and the features of routing nodes are calculated; extracting path information of information propagation according to the characteristics of the routing node, and restoring an attack path; and storing the restored path information. Therefore, according to the method, the network representation learning model is introduced, the network routing nodes are vectorized, and the characteristics of different routing nodes are considered, so that the obtained vector form can have representation and reasoning capabilities in the vector space, and then the reduced attack path is calculated by using the graph convolutional neural network; the threat perception and prediction capability of the network can be improved, and the security of the network is improved.

The invention belongs to the field of computer science, and discloses an anti-COVID-19 drug discovery method based on network characterization. The method comprises the following steps: firstly, constructing a multi-source, heterogeneous and large-scale biological medicine network by fusing a plurality of databases such as DugBank, UniProt, HPRD, SIDER, CTD, NDFRT and STRING; then, performing sequence sampling in the network in a random walk mode to form a network sequence library, and performing characterization by utilizing a deep bidirectional encoder characterization technology of Transformer to obtain a characterization vector of each node is obtained; and performing target drug interaction prediction by using an induction matrix decomposition technology so as to find a potential anti-COVID19 drug and to infer action mechanism of related drugs. According to the method, multi-source heterogeneous information and diversified data are integrated to provide multi-layer associated knowledge for drug research and development, so that the prediction precision is improved; secondly, a multi-head attention mechanism is fused through a Transformer model, the relevance between network nodes and the physical distance between the network nodes can be captured to different degrees, and then the characterization performance is improved.

The invention discloses a depression classification method based on graph embedding and a multi-modal brain network, and the method comprises the steps of modeling a brain into a brain network based on the thought of a graph theory, fusing a whole-brain function network and a structure network through a function-structure hierarchical mapping algorithm, and obtaining the brain network with the fusion of function and structure information, automatically learning the topological structure features and connection features of the brain network by using a graph embedding algorithm to obtain the vector representation of each node of the brain network, further combining into brain network representation, and finally performing depression classification based on each tested brain network representation by using a support vector machine model. According to the method, graph embedding and the multi-modal brain network are utilized, the defect of single-modal information is overcome, the feature vector suitable for a machine learning classification model is generated, the brain network level features related to depression are effectively mined, and the depression classification precision is improved.

The invention relates to a network characterization method based on an adversarial attention mechanism. The method comprises a double-mapping-function model, the first mapping function is used for distributing different weights to different node pairs through a graph attention network according to node attribute information and network topology information of real data, and an original network ismapped to a low-dimensional space to obtain more accurate low-dimensional expression of the real data; the second mapping function is node attribute information and network topology information whichare obtained by combining the obtained low-dimensional expression of the real data with disturbance to obtain noise and inputting the noise into a generator to be mapped into noise; the two functionsare used as two tuples to be input into a discriminator to be discriminated, optimization of a generator and an encoder is carried out through a result given by the discriminator, and finally low-dimensional expression which is good in robustness and capable of completely storing original network information is obtained. According to the method, the graph attention network is adopted for network representation, the correlation degree between different nodes is considered, the method is closer to the actual situation, and the effect is better.

Some embodiments include a method for determining whether to establish a connection between a computer and a network. In some embodiments, the method includes: receiving first network information from a network; determining a first trust level for the network by matching the network information to provisioning information, wherein the provisioning information indicates the first trust level and an authentication method of the network; verifying that network supports the authentication method, wherein the verifying includes exchanging information with the network; after verifying that the network supports the authentication method, establishing a connection to the network, wherein the connection provides access to an Internet.

The invention provides a network representation acquisition method based on deep learning. The method comprises the following steps: 1, obtaining a network containing node content, the network containing the node content comprises |V| nodes, selecting any one node from the |V| nodes as a current root node, performing random walk on the network containing the node content according to the current root node to obtain a content sequence and a node identification sequence; and step 2, inputting the content sequence into a deep learning model based on an attention mechanism to obtain a predicted identification vector sequence, i.e., a network representation vector, and randomly walking the network containing the node content according to the current root node to obtain a content sequence and anode identification sequence. According to the method, the research result of the deep learning technology in the machine translation direction is applied to network representation learning, and the content and the structure of the network are fused from the perspective of machine translation to obtain an appropriate network representation vector.

The invention discloses a method for evaluating opportunistic network key nodes by adopting an efficiency dependence matrix. Aiming at the dynamic change characteristics of the opportunity network, considering the global attributes and the local attributes of the nodes at the same time, the method for evaluating the opportunistic network key nodes by adopting the efficiency dependence matrix is provided. The method comprises the following steps: S1, sampling an opportunistic network, and establishing opportunistic network representation by adopting a time aggregation graph model; S2, calculating transmission efficiency and node efficiency, and sequentially constructing a transmission efficiency matrix and an efficiency dependence matrix; and S3, calculating node strength and node importance, and sorting the nodes according to the node importance. According to the method, the opportunistic network is represented as the static network through sampling and aggregation processing, and theinfluence of time on the opportunistic network structure is eliminated, and the node importance is comprehensively and accurately evaluated by comprehensively considering three indexes of transmissionefficiency, node efficiency and node strength, and the requirements of practical application are met.

A protein and nucleic acid binding site prediction method based on graph neural network characterization comprises the following steps: constructing a protein and nucleic acid interaction data set, extracting the position and feature information of each residue in the protein and the structural context thereof after sample fusion processing, and constructing graph representation of the structural context of the residues according to the graph representation. And predicting the graph representation of the to-be-predicted protein through the hierarchical graph neural network to obtain the probability that each residue is combined with DNA / RNA, thereby realizing prediction of the binding site of the protein and the nucleic acid. Key structures and feature patterns of binding sites are learned from graph representations through graph representations based on residues of structural context and a hierarchical graph neural network model.

The invention relates to a network characterization algorithm stability measurement method, which comprises the following steps of: utilizing a network characterization algorithm to keep the characteristic of a relation between network nodes, and judging the stability by using whether a nearest neighbor set of nodes in a characterization space is consistent or not; Mapping a network into a multi-dimensional continuous dense vector space by adopting a certain representation method, and calculating the first K nodes closest to each node as the basis of stability measurement by utilizing cosine similarity; a plurality of vector spaces are generated for multiple times by adopting the same representation method for one network, and the similarity of nearest neighbor sets in the vector spaces iscalculated for one node, including the coincidence rate of the nodes and the ranking information of the coincidence nodes. According to the method, the stability performance of different algorithms on different network sets can be effectively measured, so that the influence of any factor on the stability of network characterization and the influence magnitude can be further disclosed.