In this paper, we propose a novel anomaly detection method for data centers based on a combination of graphstructure and abnormal attention mechanism. The method leverages the sensor monitoring data from targetpower s...In this paper, we propose a novel anomaly detection method for data centers based on a combination of graphstructure and abnormal attention mechanism. The method leverages the sensor monitoring data from targetpower substations to construct multidimensional time series. These time series are subsequently transformed intograph structures, and corresponding adjacency matrices are obtained. By incorporating the adjacency matricesand additional weights associated with the graph structure, an aggregation matrix is derived. The aggregationmatrix is then fed into a pre-trained graph convolutional neural network (GCN) to extract graph structure features.Moreover, both themultidimensional time series segments and the graph structure features are inputted into a pretrainedanomaly detectionmodel, resulting in corresponding anomaly detection results that help identify abnormaldata. The anomaly detection model consists of a multi-level encoder-decoder module, wherein each level includesa transformer encoder and decoder based on correlation differences. The attention module in the encoding layeradopts an abnormal attention module with a dual-branch structure. Experimental results demonstrate that ourproposed method significantly improves the accuracy and stability of anomaly detection.展开更多
With the rapid development of the 5G communications,the edge intelligence enables Internet of Vehicles(IoV)to provide traffic forecasting to alleviate traffic congestion and improve quality of experience of users simu...With the rapid development of the 5G communications,the edge intelligence enables Internet of Vehicles(IoV)to provide traffic forecasting to alleviate traffic congestion and improve quality of experience of users simultaneously.To enhance the forecasting performance,a novel edge-enabled probabilistic graph structure learning model(PGSLM)is proposed,which learns the graph structure and parameters by the edge sensing information and discrete probability distribution on the edges of the traffic road network.To obtain the spatio-temporal dependencies of traffic data,the learned dynamic graphs are combined with a predefined static graph to generate the graph convolution part of the recurrent graph convolution module.During the training process,a new graph training loss is introduced,which is composed of the K nearest neighbor(KNN)graph constructed by the traffic feature tensors and the graph structure.Detailed experimental results show that,compared with existing models,the proposed PGSLM improves the traffic prediction performance in terms of average absolute error and root mean square error in IoV.展开更多
Recommendation systems play a crucial role in uncovering concealed interactions among users and items within online social networks.Recently,Graph Neural Network(GNN)-based recommendation systems exploit higher-order ...Recommendation systems play a crucial role in uncovering concealed interactions among users and items within online social networks.Recently,Graph Neural Network(GNN)-based recommendation systems exploit higher-order interactions within the user-item interaction graph,demonstrating cutting-edge performance in recommendation tasks.However,GNN-based recommendation models are susceptible to different types of noise attacks,such as deliberate perturbations or false clicks.These attacks propagate through the graph and adversely affect the robustness of recommendation results.Conventional two-stage method that purifies the graph before training the GNN model is suboptimal.To strengthen the model’s resilience to noise,we propose Graph Structure Learning for Robust Recommendation(GSLRRec),a joint learning framework that integrates graph structure learning and GNN model training for recommendation.Specifically,GSLRRec considers the graph adjacency matrix as adjustable parameters,and simultaneously optimizes both the graph structure and the representations of user/item nodes for recommendation.During the joint training process,the graph structure learning employs low-rank and sparse constraints to effectively denoise the graph.Our experiments illustrate that the simultaneous learning of both structure and GNN parameters can provide more robust recommendation results under various noise levels.展开更多
Fault diagnosis occupies a pivotal position within the domain of machine and equipment management.Existing methods,however,often exhibit limitations in their scope of application,typically focusing on specific types o...Fault diagnosis occupies a pivotal position within the domain of machine and equipment management.Existing methods,however,often exhibit limitations in their scope of application,typically focusing on specific types of signals or faults in individual mechanical components while being constrained by data types and inherent characteristics.To address the limitations of existing methods,we propose a fault diagnosis method based on graph neural networks(GNNs)embedded with multirelationships of intrinsic mode functions(MIMF).The approach introduces a novel graph topological structure constructed from the features of intrinsic mode functions(IMFs)of monitored signals and their multirelationships.Additionally,a graph-level based fault diagnosis network model is designed to enhance feature learning capabilities for graph samples and enable flexible application across diverse signal sources and devices.Experimental validation with datasets including independent vibration signals for gear fault detection,mixed vibration signals for concurrent gear and bearing faults,and pressure signals for hydraulic cylinder leakage characterization demonstrates the model's adaptability and superior diagnostic accuracy across various types of signals and mechanical systems.展开更多
Numerous works prove that existing neighbor-averaging graph neural networks(GNNs)cannot efficiently catch structure features,and many works show that injecting structure,distance,position,or spatial features can signi...Numerous works prove that existing neighbor-averaging graph neural networks(GNNs)cannot efficiently catch structure features,and many works show that injecting structure,distance,position,or spatial features can significantly improve the performance of GNNs,however,injecting high-level structure and distance into GNNs is an intuitive but untouched idea.This work sheds light on this issue and proposes a scheme to enhance graph attention networks(GATs)by encoding distance and hop-wise structure statistics.Firstly,the hop-wise structure and distributional distance information are extracted based on several hop-wise ego-nets of every target node.Secondly,the derived structure information,distance information,and intrinsic features are encoded into the same vector space and then added together to get initial embedding vectors.Thirdly,the derived embedding vectors are fed into GATs,such as GAT and adaptive graph diffusion network(AGDN)to get the soft labels.Fourthly,the soft labels are fed into correct and smooth(C&S)to conduct label propagation and get final predictions.Experiments show that the distance and hop-wise structures encoding enhanced graph attention networks(DHSEGATs)achieve a competitive result.展开更多
It is difficult to analyze semantic relations automatically, especially the semantic relations of Chinese special sentence patterns. In this paper, we apply a novel model feature structure to represent Chinese semanti...It is difficult to analyze semantic relations automatically, especially the semantic relations of Chinese special sentence patterns. In this paper, we apply a novel model feature structure to represent Chinese semantic relations, which is formalized as ＂recursive directed graph＂. We focus on Chinese special sentence patterns, including the complex noun phrase, verb-complement structure, pivotal sentences, serial verb sentence and subject-predicate predicate sentence. Feature structure facilitates a richer Chinese semantic information extraction when compared with dependency structure. The results show that using recursive directed graph is more suitable for extracting Chinese complex semantic relations.展开更多
Existing multi-view deep subspace clustering methods aim to learn a unified representation from multi-view data,while the learned representation is difficult to maintain the underlying structure hidden in the origin s...Existing multi-view deep subspace clustering methods aim to learn a unified representation from multi-view data,while the learned representation is difficult to maintain the underlying structure hidden in the origin samples,especially the high-order neighbor relationship between samples.To overcome the above challenges,this paper proposes a novel multi-order neighborhood fusion based multi-view deep subspace clustering model.We creatively integrate the multi-order proximity graph structures of different views into the self-expressive layer by a multi-order neighborhood fusion module.By this design,the multi-order Laplacian matrix supervises the learning of the view-consistent self-representation affinity matrix;then,we can obtain an optimal global affinity matrix where each connected node belongs to one cluster.In addition,the discriminative constraint between views is designed to further improve the clustering performance.A range of experiments on six public datasets demonstrates that the method performs better than other advanced multi-view clustering methods.The code is available at https://github.com/songzuolong/MNF-MDSC(accessed on 25 December 2024).展开更多
Large language models cross-domain named entity recognition task in the face of the scarcity of large language labeled data in a specific domain,due to the entity bias arising from the variation of entity information ...Large language models cross-domain named entity recognition task in the face of the scarcity of large language labeled data in a specific domain,due to the entity bias arising from the variation of entity information between different domains,which makes large language models prone to spurious correlations problems when dealing with specific domains and entities.In order to solve this problem,this paper proposes a cross-domain named entity recognition method based on causal graph structure enhancement,which captures the cross-domain invariant causal structural representations between feature representations of text sequences and annotation sequences by establishing a causal learning and intervention module,so as to improve the utilization of causal structural features by the large languagemodels in the target domains,and thus effectively alleviate the false entity bias triggered by the false relevance problem;meanwhile,through the semantic feature fusion module,the semantic information of the source and target domains is effectively combined.The results show an improvement of 2.47%and 4.12%in the political and medical domains,respectively,compared with the benchmark model,and an excellent performance in small-sample scenarios,which proves the effectiveness of causal graph structural enhancement in improving the accuracy of cross-domain entity recognition and reducing false correlations.展开更多
Meta-learning provides a framework for the possibility of mimicking artificial intelligence.How-ever,data distribution of the training set fails to be consistent with the one of the testing set as the limited domain d...Meta-learning provides a framework for the possibility of mimicking artificial intelligence.How-ever,data distribution of the training set fails to be consistent with the one of the testing set as the limited domain differences among them.These factors often result in poor generalization in existing meta-learning models.In this work,a novel smoother manifold for graph meta-learning(SGML)is proposed,which derives the similarity parameters of node features from the relationship between nodes and edges in the graph structure,and then utilizes the similarity parameters to yield smoother manifold through embedded propagation module.Smoother manifold can naturally filter out noise from the most important components when generalizing the local mapping relationship to the global.Besides suiting for generalizing on unseen low data issues,the framework is capable to easily perform transductive inference.Experimental results on MiniImageNet and TieredImageNet consistently show that applying SGML to supervised and semi-supervised classification can improve the performance in reducing the noise of domain shift representation.展开更多
Aiming at the problems of incomplete characterization of text relations,poor guidance of potential representations,and low quality of model generation in the field of controllable long text generation,this paper propo...Aiming at the problems of incomplete characterization of text relations,poor guidance of potential representations,and low quality of model generation in the field of controllable long text generation,this paper proposes a new GSPT-CVAE model(Graph Structured Processing,Single Vector,and Potential Attention Com-puting Transformer-Based Conditioned Variational Autoencoder model).The model obtains a more comprehensive representation of textual relations by graph-structured processing of the input text,and at the same time obtains a single vector representation by weighted merging of the vector sequences after graph-structured processing to get an effective potential representation.In the process of potential representation guiding text generation,the model adopts a combination of traditional embedding and potential attention calculation to give full play to the guiding role of potential representation for generating text,to improve the controllability and effectiveness of text generation.The experimental results show that the model has excellent representation learning ability and can learn rich and useful textual relationship representations.The model also achieves satisfactory results in the effectiveness and controllability of text generation and can generate long texts that match the given constraints.The ROUGE-1 F1 score of this model is 0.243,the ROUGE-2 F1 score is 0.041,the ROUGE-L F1 score is 0.22,and the PPL-Word score is 34.303,which gives the GSPT-CVAE model a certain advantage over the baseline model.Meanwhile,this paper compares this model with the state-of-the-art generative models T5,GPT-4,Llama2,and so on,and the experimental results show that the GSPT-CVAE model has a certain competitiveness.展开更多
Reinforcement learning encounters formidable challenges when tasked with intricate decision-making scenarios,primarily due to the expansive parameterized action spaces and the vastness of the corresponding policy land...Reinforcement learning encounters formidable challenges when tasked with intricate decision-making scenarios,primarily due to the expansive parameterized action spaces and the vastness of the corresponding policy landscapes.To surmount these difficulties,we devise a practical structured action graph model augmented by guiding policies that integrate trust region constraints.Based on this,we propose guided proximal policy optimization with structured action graph(GPPO-SAG),which has demonstrated pronounced efficacy in refining policy learning and enhancing performance across sophisticated tasks characterized by parameterized action spaces.Rigorous empirical evaluations of our model have been performed on comprehensive gaming platforms,including the entire suite of StarCraft II and Hearthstone,yielding exceptionally favorable outcomes.Our source code is at https://github.com/sachiel321/GPPO-SAG.展开更多
This paper investigates allocation rules in graph-structured cooperative games(hereinafter referred to as graph games)by integrating the notion of network control.A component-restricted game and the component control ...This paper investigates allocation rules in graph-structured cooperative games(hereinafter referred to as graph games)by integrating the notion of network control.A component-restricted game and the component control value are proposed through treating each connected component as a virtual player(termed a component player),under the stipulation that only coalitions attended by component players are eligible to obtain coalitional worth.The component control value initially assigns a Shapley payoff(Shapley value,SV)to every player,and the SV of each component player is subsequently redistributed equally among all original players belonging to its corresponding connected component.Thereby,an axiomatic characterization of this allocation rule is established,demonstrating that the component control value constitutes the unique solution satisfying component efficiency and component-restricted fairness in graph games.展开更多
The wide application of intelligent terminals in microgrids has fueled the surge of data amount in recent years.In real-world scenarios,microgrids must store large amounts of data efficiently while also being able to ...The wide application of intelligent terminals in microgrids has fueled the surge of data amount in recent years.In real-world scenarios,microgrids must store large amounts of data efficiently while also being able to withstand malicious cyberattacks.To meet the high hardware resource requirements,address the vulnerability to network attacks and poor reliability in the tradi-tional centralized data storage schemes,this paper proposes a secure storage management method for microgrid data that considers node trust and directed acyclic graph(DAG)consensus mechanism.Firstly,the microgrid data storage model is designed based on the edge computing technology.The blockchain,deployed on the edge computing server and combined with cloud storage,ensures reliable data storage in the microgrid.Secondly,a blockchain consen-sus algorithm based on directed acyclic graph data structure is then proposed to effectively improve the data storage timeliness and avoid disadvantages in traditional blockchain topology such as long chain construction time and low consensus efficiency.Finally,considering the tolerance differences among the candidate chain-building nodes to network attacks,a hash value update mechanism of blockchain header with node trust identification to ensure data storage security is proposed.Experimental results from the microgrid data storage platform show that the proposed method can achieve a private key update time of less than 5 milliseconds.When the number of blockchain nodes is less than 25,the blockchain construction takes no more than 80 mins,and the data throughput is close to 300 kbps.Compared with the traditional chain-topology-based consensus methods that do not consider node trust,the proposed method has higher efficiency in data storage and better resistance to network attacks.展开更多
A class of cooperative games with graph communication structure is studied in this paper by considering some important players,namely essential players.Under the assumption that only connected coalitions containing es...A class of cooperative games with graph communication structure is studied in this paper by considering some important players,namely essential players.Under the assumption that only connected coalitions containing essential players are able to cooperate and obtain their worths,the class of graph games with essential players is proposed as well as an allocation rule.The proposed value follows the spirit of the Myerson value defined by applying the Shapley value on a modified game.Three properties,feasible component efficiency,the inessential component property,and fairness,are provided to fully characterize this value,where feasible component efficiency and fairness follows the same ideas of component efficiency and fairness for classical graph games,and the inessential component property says that the total payoffs of the players in a non-feasible component is zero.Moreover,some computational aspects of the proposed value and comparisons with disjunctive permission value for games with permission structure are also studied,respectively.展开更多
A significant demand rises for energy-efficient deep neural networks to support power-limited embedding devices with successful deep learning applications in IoT and edge computing fields.An accurate energy prediction...A significant demand rises for energy-efficient deep neural networks to support power-limited embedding devices with successful deep learning applications in IoT and edge computing fields.An accurate energy prediction approach is critical to provide measurement and lead optimization direction.However,the current energy prediction approaches lack accuracy and generalization ability due to the lack of research on the neural network structure and the excessive reliance on customized training dataset.This paper presents a novel energy prediction model,NeurstrucEnergy.NeurstrucEnergy treats neural networks as directed graphs and applies a bi-directional graph neural network training on a randomly generated dataset to extract structural features for energy prediction.NeurstrucEnergy has advantages over linear approaches because the bi-directional graph neural network collects structural features from each layer's parents and children.Experimental results show that NeurstrucEnergy establishes state-of-the-art results with mean absolute percentage error of 2.60%.We also evaluate NeurstrucEnergy in a randomly generated dataset,achieving the mean absolute percentage error of 4.83%over 10 typical convolutional neural networks in recent years and 7 efficient convolutional neural networks created by neural architecture search.Our code is available at https://github.com/NEUSoftGreenAI/NeurstrucEnergy.git.展开更多
In the past decade, numerical modelling has been increasingly used for simulating the mechanical behaviour of naturally fractured rock masses. In this paper, we introduce new algorithms for spatial and temporal analys...In the past decade, numerical modelling has been increasingly used for simulating the mechanical behaviour of naturally fractured rock masses. In this paper, we introduce new algorithms for spatial and temporal analyses of newly generated fractures and blocks using an integrated discrete fracture network (DFN)-finite-discrete element method (FDEM) (DFN-FDEM) modelling approach. A fracture line calculator and analysis technique (i.e. discrete element method (DEM) fracture analysis, DEMFA) calculates the geometrical aspects of induced fractures using a dilation criterion. The resultant two-dimensional (2D) blocks are then identified and characterised using a graph structure. Block tracking trees allow track of newly generated blocks across timesteps and to analyse progressive breakage of these blocks into smaller blocks. Fracture statistics (number and total length of initial and induced fractures) are then related to the block forming processes to investigate damage evolution. The combination of various proposed methodologies together across various stages of modelling processes provides new insights to investigate the dependency of structure's resistance on the initial fracture configuration.展开更多
In feature-based visual localization for small-scale scenes,local descriptors are used to estimate the camera pose of a query image.For large and ambiguous environments,learning-based hierarchical networks that employ...In feature-based visual localization for small-scale scenes,local descriptors are used to estimate the camera pose of a query image.For large and ambiguous environments,learning-based hierarchical networks that employ local as well as global descriptors to reduce the search space of database images into a smaller set of reference views have been introduced.However,since global descriptors are generated using visual features,reference images with some of these features may be erroneously selected.In order to address this limitation,this paper proposes two clustering methods based on how often features appear as well as their covisibility.For both approaches,the scene is represented by voxels whose size and number are computed according to the size of the scene and the number of available 3Dpoints.In the first approach,a voxel-based histogram representing highly reoccurring scene regions is generated from reference images.A meanshift is then employed to group the most highly reoccurring voxels into place clusters based on their spatial proximity.In the second approach,a graph representing the covisibility-based relationship of voxels is built.Local matching is performed within the reference image clusters,and a perspective-n-point is employed to estimate the camera pose.The experimental results showed that camera pose estimation using the proposed approaches was more accurate than that of previous methods.展开更多
The information rate is an important metric of the performance of a secret-sharing scheme. In this paper we consider 272 non-isomorphic connected graph access structures with nine vertices and eight or nine edges, and...The information rate is an important metric of the performance of a secret-sharing scheme. In this paper we consider 272 non-isomorphic connected graph access structures with nine vertices and eight or nine edges, and either determine or bound the optimal information rate in each case. We obtain exact values for the optimal information rate for 231 cases and present a method that is able to derive information-theoretical upper bounds on the optimal information rate. Moreover, we apply some of the constructions to determine lower bounds on the information rate. Regarding information rate, we conclude with a full listing of the known optimal information rate （or bounds on the optimal information rate） for all 272 graphs access structures of nine participants.展开更多
The current digital educational resources are of many kinds and large quantities, to solve the problems existing in the existing dynamic resource selection methods, a dynamic resource selection method based on machine...The current digital educational resources are of many kinds and large quantities, to solve the problems existing in the existing dynamic resource selection methods, a dynamic resource selection method based on machine learning is proposed. Firstly, according to the knowledge structure and concepts of mathematical resources, combined with the basic components of dynamic mathematical resources, the knowledge structure graph of mathematical resources is constructed;according to the characteristics of mathematical resources, the interaction between users and resources is simulated, and the graph of the main body of the resources is identified, and the candidate collection of mathematical knowledge is selected;finally, according to the degree of matching between mathematical literature and the candidate collection, machine learning is utilized, and the mathematical resources are screened.展开更多
Interactive image segmentation aims at classifying the image pixels into foreground and background classes given some foreground and background markers. In this paper, we propose a novel framework for interactive imag...Interactive image segmentation aims at classifying the image pixels into foreground and background classes given some foreground and background markers. In this paper, we propose a novel framework for interactive image segmentation that builds upon graph-based manifold ranking model, a graph-based semi-supervised learning technique which can learn very smooth functions with respect to the intrinsic structure revealed by the input data. The final segmentation results are improved by overcoming two core problems of graph construction in traditional models: graph structure and graph edge weights. The user provided scribbles are treated as the must-link and must-not-link constraints. Then we model the graph as an approximatively k-regular sparse graph by integrating these constraints and our extended neighboring spatial relationships into graph structure modeling. The content and labels driven locally adaptive kernel parameter is proposed to tackle the insufficiency of previous models which usually employ a unified kernel parameter. After the graph construction,a novel three-stage strategy is proposed to get the final segmentation results. Due to the sparsity and extended neighboring relationships of our constructed graph and usage of superpixels, our model can provide nearly real-time, user scribble insensitive segmentations which are two core demands in interactive image segmentation. Last but not least, our framework is very easy to be extended to multi-label segmentation,and for some less complicated scenarios, it can even get the segmented object through single line interaction. Experimental results and comparisons with other state-of-the-art methods demonstrate that our framework can efficiently and accurately extract foreground objects from background.展开更多
基金the Science and Technology Project of China Southern Power Grid Company,Ltd.(031200KK52200003)the National Natural Science Foundation of China(Nos.62371253,52278119).
文摘In this paper, we propose a novel anomaly detection method for data centers based on a combination of graphstructure and abnormal attention mechanism. The method leverages the sensor monitoring data from targetpower substations to construct multidimensional time series. These time series are subsequently transformed intograph structures, and corresponding adjacency matrices are obtained. By incorporating the adjacency matricesand additional weights associated with the graph structure, an aggregation matrix is derived. The aggregationmatrix is then fed into a pre-trained graph convolutional neural network (GCN) to extract graph structure features.Moreover, both themultidimensional time series segments and the graph structure features are inputted into a pretrainedanomaly detectionmodel, resulting in corresponding anomaly detection results that help identify abnormaldata. The anomaly detection model consists of a multi-level encoder-decoder module, wherein each level includesa transformer encoder and decoder based on correlation differences. The attention module in the encoding layeradopts an abnormal attention module with a dual-branch structure. Experimental results demonstrate that ourproposed method significantly improves the accuracy and stability of anomaly detection.
基金supported by the project of the National Natural Science Foundation of China(No.61772562)the Knowledge Innovation Program of Wuhan-Basic Research(No.2022010801010225)the Fundamental Research Funds for the Central Universities(No.2662022YJ012)。
文摘With the rapid development of the 5G communications,the edge intelligence enables Internet of Vehicles(IoV)to provide traffic forecasting to alleviate traffic congestion and improve quality of experience of users simultaneously.To enhance the forecasting performance,a novel edge-enabled probabilistic graph structure learning model(PGSLM)is proposed,which learns the graph structure and parameters by the edge sensing information and discrete probability distribution on the edges of the traffic road network.To obtain the spatio-temporal dependencies of traffic data,the learned dynamic graphs are combined with a predefined static graph to generate the graph convolution part of the recurrent graph convolution module.During the training process,a new graph training loss is introduced,which is composed of the K nearest neighbor(KNN)graph constructed by the traffic feature tensors and the graph structure.Detailed experimental results show that,compared with existing models,the proposed PGSLM improves the traffic prediction performance in terms of average absolute error and root mean square error in IoV.
基金supported by the National Natural Science Foundation of China(Nos.62272001 and 62206002)the Anhui Provincial Natural Science Foundation(No.2208085QF195)+2 种基金the Hefei Key Common Technology Project(No.GJ2022GX15)the University Collaborative Innovation Project of Anhui Province(No.GXXT-2021-087)the Anhui Province Key Research and Development Program(No.202104a05020058).
文摘Recommendation systems play a crucial role in uncovering concealed interactions among users and items within online social networks.Recently,Graph Neural Network(GNN)-based recommendation systems exploit higher-order interactions within the user-item interaction graph,demonstrating cutting-edge performance in recommendation tasks.However,GNN-based recommendation models are susceptible to different types of noise attacks,such as deliberate perturbations or false clicks.These attacks propagate through the graph and adversely affect the robustness of recommendation results.Conventional two-stage method that purifies the graph before training the GNN model is suboptimal.To strengthen the model’s resilience to noise,we propose Graph Structure Learning for Robust Recommendation(GSLRRec),a joint learning framework that integrates graph structure learning and GNN model training for recommendation.Specifically,GSLRRec considers the graph adjacency matrix as adjustable parameters,and simultaneously optimizes both the graph structure and the representations of user/item nodes for recommendation.During the joint training process,the graph structure learning employs low-rank and sparse constraints to effectively denoise the graph.Our experiments illustrate that the simultaneous learning of both structure and GNN parameters can provide more robust recommendation results under various noise levels.
文摘Fault diagnosis occupies a pivotal position within the domain of machine and equipment management.Existing methods,however,often exhibit limitations in their scope of application,typically focusing on specific types of signals or faults in individual mechanical components while being constrained by data types and inherent characteristics.To address the limitations of existing methods,we propose a fault diagnosis method based on graph neural networks(GNNs)embedded with multirelationships of intrinsic mode functions(MIMF).The approach introduces a novel graph topological structure constructed from the features of intrinsic mode functions(IMFs)of monitored signals and their multirelationships.Additionally,a graph-level based fault diagnosis network model is designed to enhance feature learning capabilities for graph samples and enable flexible application across diverse signal sources and devices.Experimental validation with datasets including independent vibration signals for gear fault detection,mixed vibration signals for concurrent gear and bearing faults,and pressure signals for hydraulic cylinder leakage characterization demonstrates the model's adaptability and superior diagnostic accuracy across various types of signals and mechanical systems.
文摘Numerous works prove that existing neighbor-averaging graph neural networks(GNNs)cannot efficiently catch structure features,and many works show that injecting structure,distance,position,or spatial features can significantly improve the performance of GNNs,however,injecting high-level structure and distance into GNNs is an intuitive but untouched idea.This work sheds light on this issue and proposes a scheme to enhance graph attention networks(GATs)by encoding distance and hop-wise structure statistics.Firstly,the hop-wise structure and distributional distance information are extracted based on several hop-wise ego-nets of every target node.Secondly,the derived structure information,distance information,and intrinsic features are encoded into the same vector space and then added together to get initial embedding vectors.Thirdly,the derived embedding vectors are fed into GATs,such as GAT and adaptive graph diffusion network(AGDN)to get the soft labels.Fourthly,the soft labels are fed into correct and smooth(C&S)to conduct label propagation and get final predictions.Experiments show that the distance and hop-wise structures encoding enhanced graph attention networks(DHSEGATs)achieve a competitive result.
基金Supported by the National Natural Science Foundation of China(61202193,61202304)the Major Projects of Chinese National Social Science Foundation(11&ZD189)+2 种基金the Chinese Postdoctoral Science Foundation(2013M540593,2014T70722)the Accomplishments of Listed Subjects in Hubei Prime Subject Developmentthe Open Foundation of Shandong Key Lab of Language Resource Development and Application
文摘It is difficult to analyze semantic relations automatically, especially the semantic relations of Chinese special sentence patterns. In this paper, we apply a novel model feature structure to represent Chinese semantic relations, which is formalized as ＂recursive directed graph＂. We focus on Chinese special sentence patterns, including the complex noun phrase, verb-complement structure, pivotal sentences, serial verb sentence and subject-predicate predicate sentence. Feature structure facilitates a richer Chinese semantic information extraction when compared with dependency structure. The results show that using recursive directed graph is more suitable for extracting Chinese complex semantic relations.
基金supported by the National Key R&D Program of China(2023YFC3304600).
文摘Existing multi-view deep subspace clustering methods aim to learn a unified representation from multi-view data,while the learned representation is difficult to maintain the underlying structure hidden in the origin samples,especially the high-order neighbor relationship between samples.To overcome the above challenges,this paper proposes a novel multi-order neighborhood fusion based multi-view deep subspace clustering model.We creatively integrate the multi-order proximity graph structures of different views into the self-expressive layer by a multi-order neighborhood fusion module.By this design,the multi-order Laplacian matrix supervises the learning of the view-consistent self-representation affinity matrix;then,we can obtain an optimal global affinity matrix where each connected node belongs to one cluster.In addition,the discriminative constraint between views is designed to further improve the clustering performance.A range of experiments on six public datasets demonstrates that the method performs better than other advanced multi-view clustering methods.The code is available at https://github.com/songzuolong/MNF-MDSC(accessed on 25 December 2024).
基金supported by National Natural Science Foundation of China Joint Fund for Enterprise Innovation Development(U23B2029)National Natural Science Foundation of China(62076167,61772020)+1 种基金Key Scientific Research Project of Higher Education Institutions in Henan Province(24A520058,24A520060,23A520022)Postgraduate Education Reform and Quality Improvement Project of Henan Province(YJS2024AL053).
文摘Large language models cross-domain named entity recognition task in the face of the scarcity of large language labeled data in a specific domain,due to the entity bias arising from the variation of entity information between different domains,which makes large language models prone to spurious correlations problems when dealing with specific domains and entities.In order to solve this problem,this paper proposes a cross-domain named entity recognition method based on causal graph structure enhancement,which captures the cross-domain invariant causal structural representations between feature representations of text sequences and annotation sequences by establishing a causal learning and intervention module,so as to improve the utilization of causal structural features by the large languagemodels in the target domains,and thus effectively alleviate the false entity bias triggered by the false relevance problem;meanwhile,through the semantic feature fusion module,the semantic information of the source and target domains is effectively combined.The results show an improvement of 2.47%and 4.12%in the political and medical domains,respectively,compared with the benchmark model,and an excellent performance in small-sample scenarios,which proves the effectiveness of causal graph structural enhancement in improving the accuracy of cross-domain entity recognition and reducing false correlations.
基金Supported by the National Natural Science Foundation of China(No.61171131)the Key R&D Program of Shandong Province(No.YD01033)the China Scholarship Council Project(No.021608370049).
文摘Meta-learning provides a framework for the possibility of mimicking artificial intelligence.How-ever,data distribution of the training set fails to be consistent with the one of the testing set as the limited domain differences among them.These factors often result in poor generalization in existing meta-learning models.In this work,a novel smoother manifold for graph meta-learning(SGML)is proposed,which derives the similarity parameters of node features from the relationship between nodes and edges in the graph structure,and then utilizes the similarity parameters to yield smoother manifold through embedded propagation module.Smoother manifold can naturally filter out noise from the most important components when generalizing the local mapping relationship to the global.Besides suiting for generalizing on unseen low data issues,the framework is capable to easily perform transductive inference.Experimental results on MiniImageNet and TieredImageNet consistently show that applying SGML to supervised and semi-supervised classification can improve the performance in reducing the noise of domain shift representation.
文摘Aiming at the problems of incomplete characterization of text relations,poor guidance of potential representations,and low quality of model generation in the field of controllable long text generation,this paper proposes a new GSPT-CVAE model(Graph Structured Processing,Single Vector,and Potential Attention Com-puting Transformer-Based Conditioned Variational Autoencoder model).The model obtains a more comprehensive representation of textual relations by graph-structured processing of the input text,and at the same time obtains a single vector representation by weighted merging of the vector sequences after graph-structured processing to get an effective potential representation.In the process of potential representation guiding text generation,the model adopts a combination of traditional embedding and potential attention calculation to give full play to the guiding role of potential representation for generating text,to improve the controllability and effectiveness of text generation.The experimental results show that the model has excellent representation learning ability and can learn rich and useful textual relationship representations.The model also achieves satisfactory results in the effectiveness and controllability of text generation and can generate long texts that match the given constraints.The ROUGE-1 F1 score of this model is 0.243,the ROUGE-2 F1 score is 0.041,the ROUGE-L F1 score is 0.22,and the PPL-Word score is 34.303,which gives the GSPT-CVAE model a certain advantage over the baseline model.Meanwhile,this paper compares this model with the state-of-the-art generative models T5,GPT-4,Llama2,and so on,and the experimental results show that the GSPT-CVAE model has a certain competitiveness.
基金supported by National Nature Science Foundation of China(Nos.62073324,6200629,61771471 and 91748131)in part by the InnoHK Project,China.
文摘Reinforcement learning encounters formidable challenges when tasked with intricate decision-making scenarios,primarily due to the expansive parameterized action spaces and the vastness of the corresponding policy landscapes.To surmount these difficulties,we devise a practical structured action graph model augmented by guiding policies that integrate trust region constraints.Based on this,we propose guided proximal policy optimization with structured action graph(GPPO-SAG),which has demonstrated pronounced efficacy in refining policy learning and enhancing performance across sophisticated tasks characterized by parameterized action spaces.Rigorous empirical evaluations of our model have been performed on comprehensive gaming platforms,including the entire suite of StarCraft II and Hearthstone,yielding exceptionally favorable outcomes.Our source code is at https://github.com/sachiel321/GPPO-SAG.
文摘This paper investigates allocation rules in graph-structured cooperative games(hereinafter referred to as graph games)by integrating the notion of network control.A component-restricted game and the component control value are proposed through treating each connected component as a virtual player(termed a component player),under the stipulation that only coalitions attended by component players are eligible to obtain coalitional worth.The component control value initially assigns a Shapley payoff(Shapley value,SV)to every player,and the SV of each component player is subsequently redistributed equally among all original players belonging to its corresponding connected component.Thereby,an axiomatic characterization of this allocation rule is established,demonstrating that the component control value constitutes the unique solution satisfying component efficiency and component-restricted fairness in graph games.
文摘The wide application of intelligent terminals in microgrids has fueled the surge of data amount in recent years.In real-world scenarios,microgrids must store large amounts of data efficiently while also being able to withstand malicious cyberattacks.To meet the high hardware resource requirements,address the vulnerability to network attacks and poor reliability in the tradi-tional centralized data storage schemes,this paper proposes a secure storage management method for microgrid data that considers node trust and directed acyclic graph(DAG)consensus mechanism.Firstly,the microgrid data storage model is designed based on the edge computing technology.The blockchain,deployed on the edge computing server and combined with cloud storage,ensures reliable data storage in the microgrid.Secondly,a blockchain consen-sus algorithm based on directed acyclic graph data structure is then proposed to effectively improve the data storage timeliness and avoid disadvantages in traditional blockchain topology such as long chain construction time and low consensus efficiency.Finally,considering the tolerance differences among the candidate chain-building nodes to network attacks,a hash value update mechanism of blockchain header with node trust identification to ensure data storage security is proposed.Experimental results from the microgrid data storage platform show that the proposed method can achieve a private key update time of less than 5 milliseconds.When the number of blockchain nodes is less than 25,the blockchain construction takes no more than 80 mins,and the data throughput is close to 300 kbps.Compared with the traditional chain-topology-based consensus methods that do not consider node trust,the proposed method has higher efficiency in data storage and better resistance to network attacks.
基金supported by the National Natural Science Foundation of China(No.71901145)the Shanghai Planning Project of Philosophy and Social Science(No.2019EGL010).
文摘A class of cooperative games with graph communication structure is studied in this paper by considering some important players,namely essential players.Under the assumption that only connected coalitions containing essential players are able to cooperate and obtain their worths,the class of graph games with essential players is proposed as well as an allocation rule.The proposed value follows the spirit of the Myerson value defined by applying the Shapley value on a modified game.Three properties,feasible component efficiency,the inessential component property,and fairness,are provided to fully characterize this value,where feasible component efficiency and fairness follows the same ideas of component efficiency and fairness for classical graph games,and the inessential component property says that the total payoffs of the players in a non-feasible component is zero.Moreover,some computational aspects of the proposed value and comparisons with disjunctive permission value for games with permission structure are also studied,respectively.
基金supported by the Natural Science Foundation of Liaoning Province(2020-BS-054)the Fundamental Research Funds for the Central Universities(N2017005)the National Natural Science Foundation of China(62162050).
文摘A significant demand rises for energy-efficient deep neural networks to support power-limited embedding devices with successful deep learning applications in IoT and edge computing fields.An accurate energy prediction approach is critical to provide measurement and lead optimization direction.However,the current energy prediction approaches lack accuracy and generalization ability due to the lack of research on the neural network structure and the excessive reliance on customized training dataset.This paper presents a novel energy prediction model,NeurstrucEnergy.NeurstrucEnergy treats neural networks as directed graphs and applies a bi-directional graph neural network training on a randomly generated dataset to extract structural features for energy prediction.NeurstrucEnergy has advantages over linear approaches because the bi-directional graph neural network collects structural features from each layer's parents and children.Experimental results show that NeurstrucEnergy establishes state-of-the-art results with mean absolute percentage error of 2.60%.We also evaluate NeurstrucEnergy in a randomly generated dataset,achieving the mean absolute percentage error of 4.83%over 10 typical convolutional neural networks in recent years and 7 efficient convolutional neural networks created by neural architecture search.Our code is available at https://github.com/NEUSoftGreenAI/NeurstrucEnergy.git.
文摘In the past decade, numerical modelling has been increasingly used for simulating the mechanical behaviour of naturally fractured rock masses. In this paper, we introduce new algorithms for spatial and temporal analyses of newly generated fractures and blocks using an integrated discrete fracture network (DFN)-finite-discrete element method (FDEM) (DFN-FDEM) modelling approach. A fracture line calculator and analysis technique (i.e. discrete element method (DEM) fracture analysis, DEMFA) calculates the geometrical aspects of induced fractures using a dilation criterion. The resultant two-dimensional (2D) blocks are then identified and characterised using a graph structure. Block tracking trees allow track of newly generated blocks across timesteps and to analyse progressive breakage of these blocks into smaller blocks. Fracture statistics (number and total length of initial and induced fractures) are then related to the block forming processes to investigate damage evolution. The combination of various proposed methodologies together across various stages of modelling processes provides new insights to investigate the dependency of structure's resistance on the initial fracture configuration.
基金supported by the Basic Science Research Program through the National Research Foundation of Korea(NRF)funded by the Ministry of Education(NRF-2018R1D1A1B07049932).
文摘In feature-based visual localization for small-scale scenes,local descriptors are used to estimate the camera pose of a query image.For large and ambiguous environments,learning-based hierarchical networks that employ local as well as global descriptors to reduce the search space of database images into a smaller set of reference views have been introduced.However,since global descriptors are generated using visual features,reference images with some of these features may be erroneously selected.In order to address this limitation,this paper proposes two clustering methods based on how often features appear as well as their covisibility.For both approaches,the scene is represented by voxels whose size and number are computed according to the size of the scene and the number of available 3Dpoints.In the first approach,a voxel-based histogram representing highly reoccurring scene regions is generated from reference images.A meanshift is then employed to group the most highly reoccurring voxels into place clusters based on their spatial proximity.In the second approach,a graph representing the covisibility-based relationship of voxels is built.Local matching is performed within the reference image clusters,and a perspective-n-point is employed to estimate the camera pose.The experimental results showed that camera pose estimation using the proposed approaches was more accurate than that of previous methods.
基金Acknowledgements This work was supported by the National Natural Science Foundation of China （Grant No. 61373150） and the Key Technologies R ＆ D Program of Shaanxi Province （2013k0611）.
文摘The information rate is an important metric of the performance of a secret-sharing scheme. In this paper we consider 272 non-isomorphic connected graph access structures with nine vertices and eight or nine edges, and either determine or bound the optimal information rate in each case. We obtain exact values for the optimal information rate for 231 cases and present a method that is able to derive information-theoretical upper bounds on the optimal information rate. Moreover, we apply some of the constructions to determine lower bounds on the information rate. Regarding information rate, we conclude with a full listing of the known optimal information rate （or bounds on the optimal information rate） for all 272 graphs access structures of nine participants.
文摘The current digital educational resources are of many kinds and large quantities, to solve the problems existing in the existing dynamic resource selection methods, a dynamic resource selection method based on machine learning is proposed. Firstly, according to the knowledge structure and concepts of mathematical resources, combined with the basic components of dynamic mathematical resources, the knowledge structure graph of mathematical resources is constructed;according to the characteristics of mathematical resources, the interaction between users and resources is simulated, and the graph of the main body of the resources is identified, and the candidate collection of mathematical knowledge is selected;finally, according to the degree of matching between mathematical literature and the candidate collection, machine learning is utilized, and the mathematical resources are screened.
基金supported by NSFC (National Natural Science Foundation of China, No. 61272326)the research grant of University of Macao (No. MYRG202(Y1L4)-FST11-WEH)the research grant of University of Macao (No. MYRG2014-00139-FST)
文摘Interactive image segmentation aims at classifying the image pixels into foreground and background classes given some foreground and background markers. In this paper, we propose a novel framework for interactive image segmentation that builds upon graph-based manifold ranking model, a graph-based semi-supervised learning technique which can learn very smooth functions with respect to the intrinsic structure revealed by the input data. The final segmentation results are improved by overcoming two core problems of graph construction in traditional models: graph structure and graph edge weights. The user provided scribbles are treated as the must-link and must-not-link constraints. Then we model the graph as an approximatively k-regular sparse graph by integrating these constraints and our extended neighboring spatial relationships into graph structure modeling. The content and labels driven locally adaptive kernel parameter is proposed to tackle the insufficiency of previous models which usually employ a unified kernel parameter. After the graph construction,a novel three-stage strategy is proposed to get the final segmentation results. Due to the sparsity and extended neighboring relationships of our constructed graph and usage of superpixels, our model can provide nearly real-time, user scribble insensitive segmentations which are two core demands in interactive image segmentation. Last but not least, our framework is very easy to be extended to multi-label segmentation,and for some less complicated scenarios, it can even get the segmented object through single line interaction. Experimental results and comparisons with other state-of-the-art methods demonstrate that our framework can efficiently and accurately extract foreground objects from background.
