402 results about "Graph Node" patented technology

The spectral kernel machine combines kernel functions and spectral graph theory for solving problems of machine learning. The data points in the dataset are placed in the form of a matrix known as a kernel matrix, or Gram matrix, containing all pairwise kernels between the data points. The dataset is regarded as nodes of a fully connected graph. A weight equal to the kernel between the two nodes is assigned to each edge of the graph. The adjacency matrix of the graph is equivalent to the kernel matrix, also known as the Gram matrix. The eigenvectors and their corresponding eigenvalues provide information about the properties of the graph, and thus, the dataset. The second eigenvector can be thresholded to approximate the class assignment of graph nodes. Eigenvectors of the kernel matrix may be used to assign unlabeled data to clusters, merge information from labeled and unlabeled data by transduction, provide model selection information for other kernels, detect novelties or anomalies and/or clean data, and perform supervised learning tasks such as classification.

Provided is a process including: obtaining a first graph comprising nodes and edges, each of the first-graph edges linking two of the first-graph nodes and denoting semantic similarity of unstructured text in documents corresponding to the two linked first-graph nodes; for each of the first-graph nodes, selecting nodes for a second graph from attributes of the unstructured text documents to which the first-graph node corresponds, wherein the attributes are entities mentioned in the unstructured text documents, and wherein each of the second-graph nodes corresponds to a respective selected attribute; and for each pair of the second-graph nodes, determining a respective edge weight indicating similarity between a first entity corresponding to a first node of the respective pair and a second entity corresponding to a second node of the respective pair.

Computer-implemented methods, systems, and articles of manufacture for determining the importance of a data item. A method includes: (a) receiving a node graph; (b) approximating a number of neighbor nodes of a node; and (c) calculating a average shortest path length of the node to the remaining nodes using the approximation step, where this calculation demonstrates the importance of a data item represented by the node. Another method includes: (a) receiving a node graph; (b) building a decomposed line graph of the node graph; (c) calculating stationary probabilities of incident edges of a node graph node in the decomposed line graph, and (d) calculating a summation of the stationary probabilities of the incident edges associated with the node, where the summation demonstrates the importance of a data item represented by the node. Both methods have at least one step carried out using a computer device.

A method of labeling unlabeled nodes in a graph that represents objects that have an explicit structure between them. A computing device can use a labeling engine to labeled nodes in a graph that are labeled and can identify an unlabeled node in the graph that is structurally associated with the labeled nodes. The labeling engine can label the unlabeled node with the label of the labeled node based on the structural association between the unlabeled node and the labeled node.

A system, method and computer program product provides a random walk model with heterogeneous graphs to leverage multiple source data and accomplish prediction tasks. The system and method components include: 1) A heterogeneous graph formulation including heterogeneous instances of abstract objects as graph nodes and multiple relations as edges connecting those nodes. The different types of relations, such as client-vendor relation and client-product relation, are often quantified as the weights of edges connecting those entities; 2) To accomplish prediction tasks with such information, launching a multi-stage random walk model over the heterogeneous graph. The random walk within a subgraph with homogenous nodes usually produces the relevance between entities of the same type. The random walk across different type of nodes provides the prediction of decisions, such as a client purchasing a product.

The invention discloses a graph node multi-tag classification method based on depth learning. The method comprises the steps that a graph data module is loaded, and graph data are analyzed and stored in the form of a dictionary; a walk path module is generated to complete random walk in the graph data, and a generated walk path is returned; a node eigenvector module is generated, and the walk path returned by the previous step, the specified vector representing dimension and the context window size are used as input to call a word2vec algorithm to calculate the eigenvector representation of each graph node; a training data module is generated, and nodes of a certain percentage are randomly selected from all graph nodes as training node data; for each node, the eigenvector and the tag sequence corresponding to the node are taken to form a two-tuple set as a training sample; and finally, a depth confidence network model is built. The graph node multi-tag classification algorithm proposed by the invention has higher correct rate than a traditional multi-tag classification algorithm.

A graph-based iterative method is provided for selecting component modifications in an integrated circuit design that reduce the power consumption to a minimum while still meeting timing constraints. Channel-connected components are represented as nodes in a timing graph and edges in the timing graph represent directed paths. From the timing graph, a move graph is constructed containing a plurality of move nodes. Each move node represents a change to one of the components in one of the timing graph nodes. A given timing graph node can result in a plurality of move nodes. Move nodes can be merged into group nodes, and both the move nodes and group nodes are assigned a weight based on the change in power and timing effects of the associated components changes. These weights are used to select move nodes or group nodes. In general, a set of move or group nodes is selected representing the maximum cumulative weight and the components changes associated with the nodes in the set are performed on the integrated circuit design. Moves that cause timing violations are reversed. The node weights are updated following components changes and the selection of node sets is repeated iteratively until the power consumption converges to a minimum.

The invention discloses an urban traffic situation identification method based on a directed graph convolutional neural network. The method comprises the steps: carrying out the traffic situation classification of historical traffic flow information, converting an urban road network into a directed graph according to a point-edge conversion rule, and extracting a corresponding sub-graph; then, calculating the weight of a directed edge and the weight between non-directly connected nodes, standardizing the number of nodes of the subgraph, and calculating a traffic information matrix and a feature matrix of the subgraph; finally, designing a traffic directed graph convolutional neural network model, performing training and testing, using the model for classifying real-time traffic flow information to identify the real-time traffic situation of all road sections. According to the method, the incidence relation between directed road sections of different levels and different grades under the hybrid road network is fully considered, a unified standardized model input and traffic situation recognition model is designed, and good universality is achieved; moreover, the method has the characteristics of simple process, easiness in calculation, easiness in programming realization and the like, and can be suitable for complex urban road networks.

The present invention discloses a Chinese integrated entity linking method based on a graph model. An ambiguous entity in a text can be mapped into a specific entity in a real world, in order to provide aid for knowledge base expansion, information extraction and search engines. The method mainly comprises three parts of generating a candidate entity, constructing an entity indicator diagram, and disambiguating an integrated entity. For a given text, an entity referent item therein is recognized to obtain the candidate entity. The entity referent item and the candidate entity thereof are regarded as graph nodes to construct an entity referent graph. An in-degree and out-degree algorithm is applied to the entity indicator diagram for implementing disambiguation of multiple ambiguous entities in the text. The present invention does not depend on the knowledge base completely in the establishment of the entity indicator diagram, and also can implement incremental evidence mining to find evidence on an encyclopedia webpage. Dependence path analysis is employed to find the possibly related entity referent item. When the dependence path sizes of two entity referent items are within a set range, the two entity referent items are regarded as the possibly related entity referent items. Further, whether their candidate entities have relations in the real world is determined, so that the efficiency of disambiguation is greatly improved.

A block chain hybrid consensus method based on DAG algorithm comprises the following steps: 1.a client collects and verifies transactions or transaction sets; 2, selecting at least two graph nodes whose weight is high than that set value from the DAG network, verifying the transaction validity of the graph nodes or the transaction set, and proceed to the next step after the verification; 3, the client end submits a transaction or a transaction set to the DAG network; Step 4, adopting a second consensus algorithm different from the DAG consensus algorithm to carry out deterministic consensus; 5, the consensus initiator sends a graph node consensus completion message to the DAG network; Step 6: increasing the weight of the graph node in the DAG network.

The invention discloses a method for processing a virtual private cloud. The method comprises: dividing nodes with the same characteristics in cloud network topology and virtual private cloud topology into the same resource collections respectively, abstracting the resource collections into graph nodes of the cloud network topology and the virtual private cloud topology respectively, and defining and describing attribute collection of each graph node; ordering the abstracted graph nodes of the cloud network topology and the virtual private cloud topology respectively and determining adjacent matrix; operating the attribute collection of each graph node and the adjacent matrix of the cloud network topology and the virtual private cloud topology according to subgraph isomorphism algorithm, and choosing matching methods based on the obtained solution; and redistributing independent node resources of resource collection of the virtual private cloud topology and the resource collection of the cloud network topology according to the matching methods. Through the above method, the virtual private cloud customization business demand can be better met and processing efficiency is improved.

A visualization displays user designs and performance information at different levels of detail. Related register bits are combined into a metaregister and displayed as a graph node. The set of paths and associated combinatorial logic between two or more metaregisters are collapsed into a metapath and displayed as a graph connection. The set of paths associated with a metapath can be selectively revealed in response to user input. Metapaths can be annotated with performance information of its associated paths, such as timing, area, and power consumption information. The annotated performance information can represent performance information of one or more paths or aggregate attributes of the set of paths. Paths associated with control signals and finite state machines can be identified and displayed as separate graph connections.

Systems and processes providing a tool for visualizing parallel dependency graph evaluation in computer animation are provided. Runtime evaluation data of a parallel dependency graph may be collected, including the start time and stop time for each node in the graph. The visualization tool may process the data to generate performance visualizations as well as other analysis features. Performance visualizations may illustrate the level of concurrency over time during parallel dependency graph evaluation. Performance visualizations may be generated by graphing node blocks according to node start time and stop time as well as the level of concurrency at a given time to illustrate parallelism. Performance visualizations may enable character technical directors, character riggers, programmers, and other users to evaluate how well parallelism is expressed in parallel dependency graphs in computer animation.

The invention discloses a social activity recommendation method based on a heterogeneous graph model. The method comprises the steps that an offline social activity is recommended to a user based on the hobby and interest, social relation, position preference, historical behaviors, etc. of the user, and a user distribution scheme is adjusted adaptively according to the scale of the activity. The heterogeneous graph model is the core of the technology, and construction of the heterogeneous graph model mainly comprises the steps of influence factor discrimination, feature dimension reduction, heterogeneous graph node selection, node connection establishment and "virtual connection" establishment for hanging nodes. After the graph model is established, a recommendation problem is transformed into a node adjacency calculation problem in the graph model from a similarity calculation problem between the user and the activity. In a graph network, user nodes and activity nodes are connected directly or indirectly through edges, the nodes in close connection have higher relevancy, candidate activities are arranged according to a descending order according to the relevancy between the user nodes and all the activity nodes, and therefore K users with the highest values are selected to form a user recommendation list.

A method for generating mathematical analysis of a communication protocol in a vehicular communications network. The method defines features of a vehicular network, which may include a graph of a street map within a geographic area. A random geometric graph with a plurality of parameters is generated. A plurality of communications protocols on the vehicular network are defined. A communication protocol over the random geometric graph is redefined. A communication protocol's basic properties and associated features on the random geometric graph are analyzed. Results of the analysis are generated. The results of the analysis based on the random geometric graph's parameters are translated into results based on the vehicular network features. The random geometric graph with the parameters are displayed. The parameters may include: a number of graph nodes; and a probability that any two nodes are communicably connected being expressed as a function of the vehicular network features.

The invention discloses a domain knowledge pushing method based on a knowledge graph. The domain knowledge pushing method comprises the steps of collecting domain knowledge texts to construct a text knowledge base; performing semantic analysis and topic modeling on the knowledge base text; obtaining semantic distribution vectors of the nodes through domain knowledge graph embedding; establishing task context features according to the user task description and the task theme; performing entity alignment according to domain entities in the task description and the knowledge graph, performing feature extension based on graph node paths and graph node semantic distribution features, and performing task associated knowledge recall; performing text similarity calculation on the recalled text and the user task text to obtain candidate text scores; and pushing the sorting result text to the user according to the score. The text matching degree and the user experience of domain knowledge pushing are improved through node associated knowledge of the knowledge graph and a graph embedding technology.

Disclosed a multi-station scanning global point cloud registration method based on graph optimization, including acquiring multi-station original three-dimensional point cloud data; based on initial registration of targets, completing initial registration of point cloud data at adjacent stations by virtue of the target at each angle of view; calculating a point cloud overlap area at adjacent angles of view, and calculating areas of overlap regions of adjacent point cloud by a gridded sampling method; constructing a fine registration graph structure, and constructing a fine registration graph by taking point cloud data of each station as a node of the graph and taking an overlap area of the point cloud data of adjacent stations as a side of adjacent nodes of the graph structure; and based on loop closure fine registration based on graph optimization, gradually completing point cloud fine registration of the whole aircraft according to a specific closure sequence.

The invention discloses a graph processing method and device. The graph processing method comprises the steps of obtaining data of graph nodes; judging whether the data of the graph nodes change or not; if it is determined that the data of the graph nodes change, conducting redrawing processing on all the graph nodes; if it is determined that the data of the graph nodes do not change, not conducting the redrawing processing on the graph nodes. According to the graph processing method and device, the problem that CPU consumption is larger when the graph redrawing processing is carried out in the prior art is solved, and the effect that the CPU consumption is reduced when graph redrawing is carried out is achieved.