93 results about "Message propagation" patented technology

The embodiment of the invention provides a message propagation method and device, electronic equipment, a storage medium and a program product, and relates to the technical field of network communication and security. The method comprises the steps that a head end router analyzes a segment routing strategy sent by a controller to obtain a target segment list, the target segment list is composed of ordered segment identifiers, the segment identifiers in the target segment list are used for representing nodes in a network, and the target segment list is used for defining a forwarding path; the head end router generates a source address verification message based on the segment routing strategy; and the head end router forwards the source address verification message according to the forwarding path. According to the method, a source address verification message propagation mechanism based on a segment routing strategy is provided, SAV messages are guided by defining a clear path, and it is ensured that the source address verification messages are propagated efficiently, safely and accurately.

The invention discloses a multi-modal commodity recommendation method based on large language model semantic enhancement, and the method employs LightGCN to carry out message propagation, obtains user ID embedding and commodity ID embedding, employs a cross attention mechanism to learn the relevance between modals, solves a problem that the semantic space between the modals is not matched, achieves the alignment of modal features, and improves the recommendation efficiency. According to the method, the multi-modal information is matched with a semantic space of a large language model by using an adapter to enhance information association between modals, and the understanding of the multi-modal information is enhanced by using the large language model, so that the problem of unbalance of the modal information is solved, and user modal embedding and commodity modal embedding are obtained; and finally, fusing the user ID embedding and the user modal embedding to obtain a final user embedding, fusing the commodity ID embedding and the commodity modal embedding to obtain a final commodity embedding, performing inner product operation on the final user embedding and the final commodity embedding to obtain a preference score, and sorting the commodities according to the preference score to obtain a recommendation list.

The invention discloses a medical question-answering system based on a time sequence knowledge graph and a question-answering method thereof, and the system comprises a TKG construction module which is used for constructing the time sequence knowledge graph in the medical field and comprises entities, relationships and timestamp information; the hierarchical graph neural network module fused with position coding comprises a sub-graph layer and a global graph layer, the sub-graph layer is used for capturing a structural dependency relationship of concurrent facts under the same timestamp, and the global graph layer is used for capturing time correlation between cross-timestamp entities; the LLM collaborative reasoning module adopts RAG retrieval and combines the reasoning result of the TKG with an external medical knowledge base to generate an answer; the multi-mode interaction module integrates voice recognition and synthesis and supports voice questions and answers; according to the scheme, the position codes are fused into the message propagation process of the relation perception graph convolutional neural network, the distinguishing capacity of the target node for the neighbor nodes is greatly enhanced, and therefore the expression capacity of embedding of the target node is enhanced.

Techniques are described with respect to managing a distributed device message in a computing infrastructure. Such techniques are enabled through a universal interface apparatus including a plurality of serial interface adapter boards and a system-on-a-chip microcontroller. The universal interface apparatus provides a universal gateway solution between one or more component interfaces associated with a certain premises or environment and a remote system. An associated method includes deriving core message content from a distributed device message originating from a source component in a computing infrastructure, converting the derived core message content to open standard file format message content, propagating the open standard file format message content to a virtualized management system, and receiving an open standard file format message response from the virtualized management system.

The invention discloses an efficient graph comparison learning method for social recommendation. As an emerging self-supervised learning normal form, graph contrast learning is excellent in response to data sparseness and cold start due to the fact that the graph contrast learning can effectively capture similarity and heterogeneity characteristics in a graph structure, although the learning normal form achieves a good effect in a recommendation system, the graph contrast learning can be used for solving the problems of data sparseness and cold start. However, the method still faces three defects: (1) average neighbor aggregation and a non-adaptive representation reading mechanism are adopted in a message propagation process, and high-quality node representation is difficult to learn; (2) a visual angle is enhanced by depending on a random disturbance generation graph during intervention of comparative learning, which may destroy the inherent structure of graph data and further weaken the accuracy of the model; and (3) equally treating all observation samples during parameter optimization, and neglecting the difference influence of positive samples in different training stages. Specifically, aiming at the problems, the invention provides an efficient graph contrast learning method (EGCL for short). The method comprises the following steps: firstly, designing a graph adaptive propagation module, improving an information propagation rule of a graph neural network by referring to a thermonuclear thought and an attention mechanism, and realizing differentiated aggregation of neighbor nodes by adopting a learnable weight distribution strategy; secondly, designing a double contrast learning normal form which does not need graph enhancement, and realizing mutual promotion of node characterization through intra-domain contrast learning (inter-CL) and inter-domain contrast learning (inter-CL); and finally, introducing a sample weight adaptive efficient optimization algorithm, converting the training process into a double-layer optimization problem, and adaptively adjusting the contribution degree of each sample to model optimization in different stages.

The invention discloses a building operation and maintenance problem reasoning method based on a heterogeneous graph neural network. The method comprises the following steps: extracting BIM model space and component information; constructing a heterogeneous network graph; and model training and problem reasoning. Building space units are extracted from the building information model in the IFC format, space nodes are generated, components serving all the spaces are recognized, and component nodes are generated. And based on the space-component service relationship and the space-space adjacency relationship, establishing a heterogeneous network diagram containing multiple types of nodes and multiple edge relationships so as to comprehensively represent the space structure and function dependence in the building. A heterogeneous graph neural network model is used for training, and cross-space and cross-component feature aggregation and information reasoning are realized under a relation-aware message propagation mechanism. And applying a reasoning result to a building operation and maintenance stage to realize space anomaly detection and component function state diagnosis. According to the method, fusion modeling of spatial information and component information is realized, and the problem positioning precision and response efficiency in building operation and maintenance are improved.

The invention discloses a quantum error correction code decoding method, system and terminal based on momentum belief propagation, and the method comprises the steps: representing a prior message of the occurrence probability of each type of Follicle error of each quantum bit through a log-likelihood ratio, and obtaining an initialization message according to a prior probability vector, and updating messages transmitted to all adjacent variable nodes by the check node according to the initialization message to obtain an updated horizontal message, performing dynamic message updating on the posterior message of each variable node according to the updated horizontal message and the historical message to obtain an updated posterior message, and performing dynamic message updating on the posterior message of each variable node based on the updated horizontal message and the posterior message. And calculating an updated vertical message through a fixed suppression method, obtaining estimation of a real error according to a posterior message, and outputting a result. According to the method, a momentum item is introduced into message updating, a decoding algorithm based on momentum message propagation is obtained, and the decoding precision and the convergence rate of decoding on high-degenerate quantum codes are greatly improved.

The invention relates to the technical field of intelligent medicine recommendation, and provides a medicine recommendation method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the time sequence coding processing of the current treatment information of a patient, and generating a semantic vector of the current treatment information; performing multi-message propagation processing on the heterogeneous medical graph based on a graph convolutional neural network to determine knowledge memory vectors of various medical entities, and determining knowledge vectors of the current doctor seeing information based on the knowledge memory vectors; determining fusion information according to the semantic vector and the knowledge vector; and inputting the fusion information into a drug recommendation model, carrying out probability calculation processing that new drugs are used and probability calculation processing that historical drugs are reserved and used on the fusion information, and based on the determined newly-added use probability of each drug, the reserved use probability of each historical drug and an adjacent matrix of drug interaction, carrying out drug recommendation. And determining the drug combination recommended to the patient. The safety of medicine combination is improved, and the accuracy of medicine recommendation is effectively improved.

The invention discloses an improved optimization link state routing protocol for an unmanned aerial vehicle cluster under a hybrid communication architecture, the unmanned aerial vehicle cluster comprises a plurality of unmanned aerial vehicle nodes, and each node is simultaneously equipped with an omnidirectional RF transceiver antenna and a directional FSO transceiver; in the routing process, the RF link is used for controlling stable distribution of messages, and the FSO link is used for high-speed transmission of service data; a multi-index multi-point relay node comprehensive selection algorithm is designed, and message propagation is optimized and controlled by combining node coverage, average two-hop link stability and relay load indexes; establishing a real-time link quality prediction mechanism based on unscented Kalman filtering, and generating a routing path of stability perception; and a routing compression method adaptive to heterogeneous link propagation characteristics is developed, and redundant hops are eliminated. Through the improved OLSR protocol, the problems of topological response lagging, insufficient link stability and low heterogeneous network efficiency of a traditional routing protocol under a hybrid communication architecture are effectively solved.

The invention discloses a Doppler frequency shift-based high-precision time synchronization method, which comprises the following steps that: a common node sends a time synchronization request message to a time reference node, and when the time reference node receives the time synchronization request message, the time reference node records the time when the time synchronization request message reaches an air interface and estimates the Doppler frequency shift of the time synchronization request message; the time reference node replies a time synchronization response message, and when the common node receives the time synchronization response message replied by the time reference node, the time when the time synchronization response message reaches an air interface is recorded, and meanwhile, the Doppler frequency shift of the time synchronization response message is estimated; and the common node solves the time synchronization error between the common node and the time reference node by using the Doppler frequency shift of the time synchronization request message and the time synchronization response message and the transmission delay of the time synchronization request message and the time synchronization response message, thereby realizing time synchronization. The problem of asymmetry of message propagation paths is solved, and the time synchronization precision of the mobile ad hoc network is improved.

The invention discloses a message control method, a product, electronic equipment and a medium, and relates to the technical field of data processing. The purpose of controlling message propagation is achieved by removing nodes in the message propagation process, and in the node removal process, a first preset number of selected root nodes are removed from a tree structure, and a second preset number of selected child nodes are selected from sub-trees corresponding to the root nodes. The removed nodes are selected by taking the root nodes and the sub-trees corresponding to the root nodes as units, that is, the removed nodes are determined by taking the local sub-tree structures as units, and the local sub-tree structures are divided from the whole network structure, so that the scale of the local sub-tree structures is smaller than that of the whole network, and therefore, the removed nodes can be determined by taking the root nodes and the sub-trees corresponding to the root nodes as units. Compared with a mode of analyzing by taking the overall information of the whole network as a unit, the method provided by the invention has the advantages that the calculation complexity is reduced, the response speed is high, the efficiency of determining the removal mode is improved, and the efficiency of controlling message propagation is improved.

The application discloses a method for predicting disease lncRNA based on layer refinement graph convolutional neural network, comprising the following steps: fusing various similarity information of nodes; constructing a heterogeneous network; using a graph convolutional neural network with a layered refinement mechanism to learn the feature aggregation information of each node in the heterogeneous network; integrating the feature information of each hidden layer to obtain the final feature expression of the node; applying matrix transpose multiplication to the final feature expression matrix to obtain the prediction correlation score of the lncRNA-disease pair. The layered refinement mechanism is integrated into the graph convolutional neural network, so that the similarity between each hidden layer and the initial feature layer of the node can be fully considered in the message propagation process, the weight of the layer more similar to the initial layer can be amplified, and the weight of the layer less similar to the initial layer can be reduced, so that the common over-smoothing problem in the deep neural network is solved.

The invention belongs to artificial intelligence, and particularly relates to a graph neural network confrontation defense method based on attribute enhancement PPR and self-loop weight adjustment, which comprises the following steps: a node stops at a current node or walks to a neighbor node at a certain probability, and executes topology migration or attribute migration at a certain probability during walking, calculating similarity among nodes by considering topology and attribute characteristics of multi-hop high-order information; utilizing similarity scores to cut off nodes with insufficient similarity for reconstruction, endowing corresponding connecting edges with weight values based on the similarity scores among the nodes, and reconstructing an adjacent matrix; in the message aggregation process, the proportion of the node itself and the neighbor node thereof in the final message is dynamically adjusted based on the out-degree of the target node, so that the proportion of the neighbor node is large when the number of neighbors is large, and the proportion of the node itself is large when the number of neighbors is small; performing message propagation through the reconstructed adjacent matrix; through iterative message aggregation and propagation, training a robust graph neural network with confrontation and defense capabilities; according to the invention, the robustness of the graph neural network is improved.

The application discloses a student programming answer prediction method combining semantic understanding and structured modeling, collects student programming homework data and pre-processes the data to obtain data samples, translates the data samples to obtain an English feature set; uses low-rank adaptation to fine-tune a pre-trained semantic understanding model, constructs input features, and predicts an answer correctness probability; performs graph neural network structured modeling based on the association relationship between problems and concepts, performs message propagation and feature updating to obtain an embedding set, inputs the embedding set and student embedding into a prediction layer to obtain a probability of answering a question correctly; and inputs the probability of answering a question correctly and the probability of answering a question correctly after fusion into a multilayer perceptron for nonlinear mapping to predict the probability of answering a question correctly. The application solves the problems of difficulty in effectively processing non-standardized codes submitted by students, insufficient semantic understanding of question texts and low prediction accuracy in programming knowledge tracking, and provides a scientific basis for personalized teaching and learning resource scheduling.

This invention provides a method and device for automatically detecting eye contact. The method includes: Step 1, extracting features from a node image to obtain a feature image; Step 2, updating the structure of the feature image; Step 3, updating the network nodes of the structure-updated feature image; Step 4, classifying and outputting eye contact behaviors. This invention improves the accuracy of predicting eye contact social behavior by extracting context-specific information about a person's gaze behavior, network topology information, and synchronization timing information between gaze events, and assigning different attention coefficients to the message propagation path of the graph network.

The invention discloses a hierarchical text classification method and system for ensuring node label consistency, and relates to the technical field of natural language processing, and the method comprises the steps: extracting context features of each news article through a text encoder; inputting the context features into a multi-label classifier, and outputting a preliminary classification probability corresponding to a label of each context feature; each label serves as a node, and the preliminary classification probability of each node is reconstructed into the confidence coefficient of the node through a reconstruction function; and performing message generation based on the child nodes, the father node and the confidence coefficient, aggregating the generated messages until message propagation converges, performing node state updating through the converged aggregated messages, and outputting a final multi-label classification result of the news article. According to the method, the label consistency in the classification task is realized through a message passing mechanism constrained by a logic rule, and meanwhile, the accuracy of a classification result and the integrity of a hierarchical structure are ensured.

The invention discloses an Ethereum malicious sample detection method and system for homogeneous enhanced compression modeling, and belongs to the technical field of graph artificial intelligence and block chain security. According to the technical scheme, the method comprises the steps that training label account interaction information is extracted to construct an initial Ethereum interaction graph; retaining key nodes and edges based on message propagation to generate a simplified graph; dividing target nodes, bridge nodes and background nodes; aggregating background node features to a target node, deleting the node, and aggregating bridge nodes of the same kind to form a new bridge node; deleting new bridge nodes which have the same features and have intersections of source bridge nodes to generate a compressed graph; pseudo labels are generated through a full connection layer, the same proton graph is divided, and fusion output is carried out after message transmission in the sub-graph; and training the model by combining pseudo label loss and classification loss. According to the method, the calculation complexity is reduced through graph compression, similar information propagation is enhanced by using a pseudo tag and a proton graph, and the malicious behavior detection accuracy and efficiency are improved.

The application discloses a sequence recommendation method based on meta-path neighborhood target generalization. The method comprises the following steps: modeling items and their co-occurrence as an item-item graph according to the historical behavior sequence of a user, constructing an item-label graph according to the corresponding relationship between each item and a label, and constructing a meta-path for capturing different composite relationships between items or labels; encoding the historical interaction information of the user by using a heterogeneous graph, and constructing the relationship between items and items, items and labels, and adopting a message propagation mechanism and a message aggregation mechanism to fuse different types of nodes and their relationships; carrying out embedding learning training on the heterogeneous graph to obtain a trained sequence recommendation model, in the training, for each target prediction, generalizing potential item targets to assist in training, and inputting the vector representation of the item node into a conversion layer. The application can improve the accuracy of recommended items, and explore potential interest items of the user in addition to historical interaction behaviors.

The present invention belongs to the field of computer image processing technology, and specifically discloses a multi-label oil leakage identification method for power equipment based on scene perception. The method is based on the characteristic that the co-occurrence relationship of different types of labels in an image is closely related to the image scene, and proposes a multi-label oil leakage identification model based on scene perception. The model realizes fine-grained modeling of the co-occurrence relationship of labels at the scene level through scene category detection, label co-occurrence modeling, and scene classification loss function calculation. Then, a label graph is constructed, and message propagation and feature update are performed on the label graph, so that there are more feature interactions between labels with high co-occurrence probabilities in different scenes, thereby improving each other's visual representation and promoting the identification of oil leakage labels. The multi-label oil leakage identification method for power equipment based on scene perception proposed by the present invention can not only identify the oil-stained area on the power equipment, but also accurately judge different types of oil leakage, thereby realizing multi-label oil leakage identification.

The application discloses a multi-tenant message isolation method, system and device based on a logical message domain. A message broker node maintains a mapping relationship between a client and a logical message domain, an output propagation strategy and a receiving propagation strategy. The logical message domain is internal control metadata distinguished from a message topic namespace. After receiving a publishing request, a logical message domain to which a publishing client belongs is determined, a propagation tag set is generated according to the output propagation strategy, and a message propagation context containing a message topic, a message domain identifier and the propagation tag set is established. Topic matching is performed on the message topic and a subscription topic, and message propagation visibility is calculated according to the propagation tag set and a receiving propagation strategy corresponding to a logical message domain to which a subscription client belongs. Message delivery is performed when topic matching, message visibility and subscription validity are met. The application realizes multi-tenant logical isolation and controlled cross-domain message sharing.

The application relates to a tracing method for latent period asymptomatic transmission based on an SAIR model, which comprises the following steps: S1, determining a single initial transmission source node in an interpersonal relationship complex network, and transmitting according to the source node according to an SAIR model until a moment t0; S2, randomly selecting a certain number of nodes as observation nodes, and knowing the state of the observation nodes at the moment; S3, screening a candidate source node set in the complex network by using a reverse propagation algorithm according to the state of the partial observation nodes, and the hop distance and the effective distance, and the number of nodes is determined by a sampling method; S4, constructing a dynamic message transmission equation according to the SAIR model, and calculating the probability of the candidate source nodes being in different states at a moment Delta t by using the partial observation nodes; and S5, calculating the likelihood probability of the nodes in the candidate source node set becoming sources by using an average field theory, and the source is the node with the maximum probability. According to the application, the spread of infectious diseases of patients with latent period and asymptomatic can be accurately reflected, and the problem of tracing the source of infection can be solved.

The invention discloses a signal detection method of an orthogonal time-frequency space modulation system, which is suitable for a high-speed mobile communication scene with a sparse multipath structure. According to the method, on the basis of a traditional message propagation algorithm, a layered message propagation mechanism is provided, observation nodes corresponding to received signals are divided into a plurality of layers, and message updating is executed layer by layer in each iteration: the observation nodes in each layer transmit interference term mean value and variance information to variable nodes; and the variable node immediately returns the updated symbol probability quality function. The transmission efficiency of the information in the factor graph is remarkably improved by allowing the updated information in the same round of iteration to participate in the next layer of calculation in advance. Compared with the original synchronous iteration mode, the method provided by the invention has the advantages that the convergence speed is accelerated and the overall calculation complexity is reduced while the similar bit error rate performance is maintained, and the method is suitable for a high-dynamic and low-delay wireless communication system.

The application provides a dynamic granular ball graph neural network working method for user and item recommendation, and belongs to the technical fields of recommendation systems and graph neural networks. The method comprises the following steps: S1, constructing user-side and item-side granular balls in a low-dimensional spectral space, updating the granular balls according to the node representation of each propagation layer, calculating the membership of nodes to the granular balls by using a radius-normalized distance, and defining the collaborative purity according to the interaction distribution of the granular balls on the opposite granular balls; S2, coupling fine-grained interaction modeling and coarse-grained collaborative abstraction, and performing message propagation along two complementary paths; S3, performing bilateral degree perception gating, and adaptively balancing the contributions of the two propagation paths for each node; and S4, aggregating the representations of each layer by weighted summation to obtain the final embedding of the user side and the item side, finally adopting a Bayesian personalized ranking loss to optimize the candidate item score, so as to play a good recommendation role in ranking the user and the item.

The invention relates to a hierarchical broadcasting method and device based on statistical adaptive threshold and scale constraint, and the method comprises the steps: calculating a local delay threshold and a domain scale upper limit of each network node based on the statistical analysis of local network delay; forming a plurality of target broadcast domains by adopting a bidirectional delay verification mechanism based on a local delay threshold and a domain scale upper limit; and on the basis of the formed target broadcast domain, intra-domain message propagation and cross-domain message forwarding are respectively carried out to realize layered broadcast. Through the method and the device, statistical analysis of local network delay can be performed, and the local delay threshold and the domain scale constraint of each network node are dynamically calculated, so that a self-organizing domain is formed on the basis of the local delay threshold and the domain scale constraint to perform layered broadcast; self-adaptive layered broadcasting capable of adapting to network topology dynamic change and node delay distribution difference is realized, and the problem of lack of a self-adaptive layered broadcasting mechanism is solved.

The present invention relates to a method and system for detecting abnormal user behavior in a heterogeneous information network based on attention. First, the historical interaction data of a heterogeneous information network over a certain period of time is converted into graph data, where each node in the graph data represents a constituent object of the heterogeneous information network, and the edges in the graph data reflect the connections between the constituent objects of the heterogeneous information network. Then, based on a graph neural network, an objective function of the user abnormal behavior detection model is constructed. The model aggregates the neighbor information of nodes through attention and derives an inter-layer propagation formula for node attribute representations. Finally, the gradient of each node attribute representation is updated until all node attribute representations converge to obtain each node attribute representation. The node attribute representations are compressed by a multi-layer perceptron into a one-dimensional column vector vertically stacked by the prediction labels of each node, which is the user abnormal behavior detection result. This method uses the attention mechanism to automatically capture all meta-path information in the network during the message propagation process, improving the accuracy of user abnormal behavior detection.

Methods and systems describe multicasting the messages over a broadcast medium (e.g., a TV network). In particular, the methods and systems recite the use of Advanced Television Systems Committee (“ATSC”) 3.0, a new broadcast television transmission standard in the United States. Using ATSC 3.0, the methods and systems can disseminate messages to all the electrical equipment.

This application discloses a high-precision time synchronization method based on Doppler frequency shift, comprising: a regular node sending a time synchronization request message to a time reference node; upon receiving the time synchronization request message, the time reference node recording the arrival time of the time synchronization request message at the air interface and estimating the Doppler frequency shift of the time synchronization request message; the time reference node replying with a time synchronization response message; upon receiving the time synchronization response message from the time reference node, the regular node recording the arrival time of the time synchronization response message at the air interface and estimating the Doppler frequency shift of the time synchronization response message; the regular node using the Doppler frequency shifts of the time synchronization request message and the time synchronization response message, as well as the transmission delays of the time synchronization request message and the time synchronization response message, to calculate the time synchronization error between the regular node and the time reference node, thereby achieving time synchronization. This solves the problem of message propagation path asymmetry and improves the time synchronization accuracy of mobile ad hoc networks.

The application discloses a kind of wisdom agricultural trade information management system based on multi-agent cooperation, comprising: data acquisition processing module, for collecting business data and pre-processing;Multi-agent construction module, for building multi-agent, and setting state space, action space and interaction space;Local state generation module, for structure coding, and cutting, while performing dimension and order constraint;Cooperative decision module, for generating action intention vector, performing message propagation and multi-round merging processing, and performing dependency analysis and matrix assembly on the merging result;Instruction generation module, for hierarchical analysis, behavior binding and structure rearrangement on joint action matrix, generating instructions and issuing to corresponding terminal;Strategy updating module, for collecting feedback data, and iteratively updating action generation parameters.The application can realize the cooperative processing and dynamic management of multi-source business in the wisdom agricultural trade scene, improve the information management precision.

The present invention relates to a Beidou B2b-PPP user distance precision factor refinement and user positioning random model optimization method; it belongs to the field of satellite navigation technology, and the method includes: a system end receives observation data of multiple Beidou / GNSS reference stations and B2b correction numbers and broadcast ephemeris, re-estimates Beidou B2b-PPP user distance precision factor URA in real time, and refines the user distance precision factor URA; the system end encodes the refined user distance precision factor URA and broadcasts it to the user end; the user end establishes a Beidou B2b positioning function model for B2b precise single point positioning; in the optimized random model corresponding to the positioning function model, the user uses the decoded user distance precision factor URA to calculate the observation noise. The present invention solves the problem of low accuracy of URA broadcast by B2b through re-estimation, broadcasts re-estimated URA through short messages, and does not occupy B2b signal bandwidth; the URA index is used for determining the B2b-PPP random model, and the user positioning random model is refined to improve positioning accuracy.

The application discloses a time sequence perceived heterogeneous graph neural rumor detection model. In recent years, the development of online social media greatly accelerates the breeding and spread of rumors, and the harmfulness of rumors makes the automatic detection technology of rumors widely concerned by researchers. The application simultaneously considers the global structural relationship between events and the time sequence relationship of internal message transmission of the events, models the two kinds of relationships together by taking a heterogeneous graph as a carrier, and proposes a new time sequence perceived heterogeneous graph neural rumor detection model. The model captures the time sequence relationship between the internal forwarding (or comment) posts of the events by using a time sequence perceived self-attention mechanism, fuses the forwarding (or comment) posts with time sequence information with source posts, and obtains local time sequence representations of the events. Then, the model captures the global structural relationship between the events by using an element-level attention mechanism, learns global structural representations of the events, and finally fuses the two kinds of representations for detecting rumors.