Search

In this paper, we study the graph classification problem in vertex-labeled graphs. Our main goal is to classify graphs by comparing their higher-order structures thanks to heat diffusion on their simplices. We first represent vertex-labeled graphs as simplex-weighted super-graphs. We then define the diffusion Fréchet function over their simplices to encode the higher-order network topology and finally reach our goal by combining the function values with machine learning algorithms. Our experiments on real-world bioinformatics networks show that using diffusion Fréchet function on simplices is promising in graph classification and more effective than the baseline methods. To the best of our knowledge, this is the first study in the literature using heat diffusion on higher-dimensional simplices in a graph mining problem. We believe that our method can be extended to different graph mining domains, beyond the graph classification problem.

The last half-decade has seen a surge in deep learning research on irregular domains and efforts to extend convolutional neural networks (CNNs) to work on irregularly structured data. The graph has emerged as a particularly useful geometrical object in deep learning, able to represent a variety of irregular domains well. Graphs can represent various complex systems, from molecular structure, to computer and social and traffic networks. Consequent on the extension of CNNs to graphs, a great amount of research has been published that improves the inferential power and computational efficiency of graph-based convolutional neural networks (GCNNs). The research is incipient, however, and our understanding is relatively rudimentary. The majority of GCNNs are designed to operate with certain properties. In this survey we review of the state of graph representation learning from the perspective of deep learning. We consider challenges in graph deep learning that have been neglected in the majority of work, largely because of the numerous theoretical difficulties they present. We identify four major challenges in graph deep learning: dynamic and evolving graphs, learning with edge signals and information, graph estimation, and the generalization of graph models. For each problem we discuss the theoretical and practical issues, survey the relevant research, while highlighting the limitations of the state of the art. Advances on these challenges would permit GCNNs to be applied to wider range of domains, in situations where graph models have previously been limited owing to the obstructions to applying a model owing to the domains’ natures.

While graph drawing focuses more on the aesthetic representation of node-link diagrams, graph visualization takes into account other visual metaphors making them useful for graph exploration tasks in information visualization and visual analytics. Although there are aesthetic graph drawing criteria that describe how a graph should be presented to make it faster and more reliably explorable, many controlled and uncontrolled empirical user studies flourished over the past years. The goal of them is to uncover how well the human user performs graph-specific tasks, in many cases compared to previously designed graph visualizations. Due to the fact that many parameters in a graph dataset as well as the visual representation of them might be varied and many user studies have been conducted in this space, a state-of-the-art survey is needed to understand evaluation results and findings to inform the future design, research, and application of graph visualizations. In this article, we classify the present literature on the topmost level into graph interpretation, graph memorability, and graph creation where the users with their tasks stand in focus of the evaluation, not the computational aspects. As another outcome of this work, we identify the white spots in this field and sketch ideas for future research directions.

A maximal clique ( ${MC}$ ) is a complete subgraph satisfying that no other cliques can take it as their proper subgraph. Given an uncertain graph, the $top\text{-}K\,\,{MC}\text{s}$ enumeration problem studies how to return $k {MC}\text{s}$ with the highest rank value. Existing algorithms rank ${MC}\text{s}$ according to their probabilities, thus usually return ${MC}\text{s}$ with higher probabilities but less number of vertices, and fail to return large ${MC}\text{s}$ that convey more useful information. Considering this problem, this paper studies the problem of enumerating $top\text{-}K\,\,{MC}\text{s}$ . Our approach returns $k {MC}\text{s}$ with the most number of vertices satisfying that their probabilities $\geq \alpha $ , where each ${MC}$ is called an $\alpha $ - ${MC}$ , and computing $k$ largest $\alpha $ - ${MC}\text{s}$ is called as $(k,\alpha)$ - ${MC}\text{s}$ . We propose an efficient $(k,\alpha)$ - ${MC}\text{s}$ enumeration algorithm, $Top\text{-}KMC$ , which works in three steps, including partition, enumeration and verification. Here, partition means that we compute the set $\mathcal {M}$ of all ${MC}\text{s}$ without considering the probability information, as if the graph is partitioned into a set of subgraphs. Enumeration means that we compute $\alpha $ - ${MC}\text{s}$ from each ${MC}$ of $\mathcal {M}$ . As each such subgraph is an ${MC}$ , the cost of computing common neighbors for finding $\alpha $ - ${MC}\text{s}$ can be reduced. Verification means that we need to verify whether an $\alpha $ - ${MC}$ is a subgraph of another $\alpha $ - ${MC}$ . If not, it is an $\alpha $ - ${MC}$ ; otherwise, it is a useless $\alpha $ - ${MC}$ and should be removed. We further propose an optimized algorithm $Top\text{-}KMC+$ to reduce both time and space by merging the above three steps into a whole step. The experimental results on real datasets show that both $Top\text{-}KMC$ and $Top\text{-}KMC+$ can return $k$ largest $\alpha $ - ${MC}\text{s}$ efficiently.

If the three-dimension data of batch process are unfolded the two-dimension data, some important information would lose, and outliers such as noise would lead to poor monitoring results. Therefore, a Markov chain neighborhood sparse preserving graph embedding algorithm based on tensor factorization (TMNSPGE) is proposed. Firstly, tensor factorization is used to directly process the three-dimension data in batch process, which can avoid the information loss. Secondly, by using the neighborhood preserving embedding algorithm and sparse manifold coding, the local linear relationship and local sparse manifold structure of data are preserved. On this basis, Markov chain analysis is introduced to construct a similar graph to make the data after dimensionality reduction have a certain probability interpretation. Finally, the statistics and control limits are determined to realize process monitoring. Numerical example and penicillin fermentation simulation process prove the effectiveness of TMNSPGE algorithm in batch process monitoring.

The web graph can be used to get insight into the internal structure and connectivity of the Tor dark web. This paper analyzes the internal structure of the Tor dark web graph and examines the presence of bow-tie structure as found in the World Wide Web. The web graph is generated from the data collected by the Python crawler customized to scrape data from the Tor dark web. Each of the nodes in the graph represents an individual Tor hidden service, and an edge denotes the hyperlink from one hidden service to the other. Various graph metrics are then computed and analyzed for both directed and undirected graphs using the Python NetworkX package. It was found that most of the nodes of the graph have in-degree and out-degree less than ten. The presence of power-law in degree distribution could neither be confirmed nor denied. The Tor web graph is sparse with a few connected pairs of nodes. Like the surface web, the dark web can also be decomposed into a bow-tie structure though with small component sizes. Several important and well-known websites on the surface web have incoming links from the dark web. Moreover, the Tor network also shows the characteristics of small-world and scale-free networks.

Knowledge Graphs (KGs), as one of the key trends which are driving the next wave of technologies, have now become a new form of knowledge representation, and a cornerstone for several applications from generic to specific industrial use cases. However, in some specific domains such as law enforcement, a real and large domain-oriented KG is often unavailable due to data privacy concerns. In such domains it is necessary to generate a synthetic KG which mimics the properties of a real KG in the domain. Although during the last two decades, a variety of graph data generators has been proposed to achieve the generation of different kinds of networks, the state-of-the-art synthetic graph data generators are not feasible to generate a realistic and synthetic KGs because KGs always contain data characteristics with specified semantics. In this work, we propose a schema-driven synthetic KG generation approach with extended graph differential dependencies ( $GDD^{x}$ ), which is an extension of the recently developed graph entity/differential dependencies that represent formal constraints for graph data to enable the generation of desired graph patterns in synthetic KG. Next, we develop an effective KG generation algorithm that employs the schema and the pre-defined $GDD^{x}\text{s}$ . Finally, we evaluate our synthetic KG generator and compare with several state-of-the-art synthetic graph generators. The results from the experiments show that our KG generation method can generate KGs that exhibit the desired graph patterns, node attributes and degree distributions associated with each entity type in the graph’s schema.

In the natural and social systems of the real world, various network can be seen everywhere. The world where people live can be seen as a combination of network with different dimensions. Link prediction formalizes the interaction behavior between people. Traditional link prediction methods mainly study the user behavior of static social network. This article studied the dynamic graph representation learning so as to put forward an improved link prediction model in dynamic social network. Besides, the interactions in the real world can be multiple, links at different moments may have different meanings. The proposed model firstly solved the problem of link prediction on multiple kinds of edges. The whole embedding of each node is separated into two parts, basic embedding and edge embedding. Then the proposed model selected time slices for dynamic social network to get the graph embeddings in different snapshots. What’s more, the $t+1$ time step embedding vector was used to validate $t$ time step prediction effect and the proposed model performed better than traditional graph representation learning methods.

Semantic information from images can be used to improve the performance of deep learning methods in recognizing human emotions. In this paper, we propose a novel framework based on the graph convolutional network for emotion recognition by utilizing the semantic relationships of different regions. First, we extract the salient image regions within video frame clips by using the bottom-up attention module to construct the node features of a graph. Then, we build the graphs containing the node features and the semantic correlations of nodes by using the graph convolutional network. For refinement, each node feature of graph vectors is enhanced via a gated recurrent unit consisting of gate and memory units to remove redundant feature information. Experimental results show that our proposed method achieves superior performance over state-of-the-art approaches for the emotion recognition on the CEAR and AFEW datasets.

In recent years, skeleton-based action recognition, graph convolutional networks, have achieved remarkable performance. In these existing works, the features of all nodes in the neighbor set are aggregated into the updated features of the root node, while these features are located in the same feature channel determined by the same $1\times 1$ convolution filter. This may not be optimal for capturing the features of spatial dimensions among adjacent vertices effectively. Besides, the effect of feature channels that are independent of the current action on the performance of the model is rarely investigated in existing methods. In this paper, we propose cross-channel graph convolutional networks for skeleton-based action recognition. The features fusion mechanism in our network is cross-channel, i.e, the updated feature of the root node is derived from different feature channels. Because different feature channels come from different $1 \times 1$ convolution filters, the cross-channel fusion mechanism significantly improves the ability of the model to capture local features among adjacent vertices. Moreover, by introducing a channel attention mechanism to our model, we suppress the influence of feature channels unrelated to action recognition on model performance, which improves the robustness of the model against the feature channels independent of the current action. Extensive experiments on the two large-scale datasets, NTU-RGB+D and Kinetics-Skeleton, demonstrate that the performance of our model exceeds the current mainstream methods.

Graph Neural Networks (GNNs) have been applied in many fields of semi-supervised node classification for non-Euclidean data. However, some GNNs cannot make good use of positive information brought by nodes which are far away from each central node for aggregation operations. These remote nodes with positive information can enhance the representation of the central node. Some GNNs also ignore rich structure information around each central node's surroundings or entire network. Besides, most of GNNs have a fixed architecture and cannot change their components to adapt to different tasks. In this article, we propose a semi-supervised learning platform ATPGNN with three variable components to overcome the above shortcomings. This novel model can fully adapt to different tasks by changing its components and support inductive learning. The key idea is that we first create a high-order topology graph, which is from similarity of node structure information. Specifically, we reconstruct the relationships between nodes in a potential space obtained by network embedding in graph. Second, we introduce graph representation learning methods to extract representation information of remote nodes on the high-order topology graph. Third, we use some network embedding methods to get graph structure information of each node. Finally, we combine the representation information of remote nodes, graph structure information and feature for each node by attention mechanism, and apply them to learning node representation in graph. Extensive experiments on real attributed networks demonstrate the superiority of the proposed model against traditional GNNs.

Traditional Chinese Medicine (TCM) clinical informatization focuses on serving user-oriented health knowledge and facilitating online diagnosis. Regularities are hidden in clinical knowledge play a significant role in the improvement of the TCM informatization service. However, many regularities can hardly be discovered because of specific data-challenges in TCM prescriptions at present. Therefore, in this article, we propose an end-to-end model, called Semantic-aware Graph Convolutional Networks (SaGCN) model, to learn the latent regularities in three steps: (1) We first construct a heterogeneous graph based on prescriptions; (2) We stack Semantic-aware graph convolution to learn effective low-dimensional representations of nodes by meta-graphs and self-attention; (3) With the learned representations, we can detect regularities accurately by clustering and linked prediction. To the best of our knowledge, this is the first study to use metagraph and graph convolutional networks for modeling TCM clinical data and diagnosis prediction. Experimental results on three real datasets demonstrate SaGCN outperforms the state-of-the-art models for clinical auxiliary diagnosis and treatment.

Unsupervised person re-identification (re-id) is an effective analysis for video surveillance in practice, which can train a pedestrian matching model without any annotations, and it is easy to deploy in unseen camera scenarios. The most challenging problem in unsupervised re-id task is the huge distribution-gap among different camera views, and the intrinsic correlations in unlabeled identities are also complicated to sufficiently explored. This paper proposes a Multi-order Cross-view Graph adversarial Network (MCGN) to bridge the cross-view distribution-gap, and mine the inherent discriminative information by multi-order triplet correlations. Specifically, MCGN firstly exploits graph representations by a cross-view graph convolutional network according to intra-view and inter-view graph structure, and then encodes each pedestrian image into a view-shared feature space, which is iteratively trained by a graph generative adversarial learning strategy to deeply bridge the distribution-gap. Finally, this paper proposes a multi-order discriminative learning module for composing reasonable triplet samples according to multi-order similarity correlations among unlabeled pedestrian images. Furthermore, sufficient experiments are conducted in two large scale person re-id datasets (Market-1501 and DukeMTMC-reID). The comparison to state-of-the-art methods and ablation study demonstrate the superiority of MCGN and the contribution of each module proposed in this paper.

The line segments of multiple (and nearly parallel) planar laser sources appearing in monocular video are objectively associated to their respective laser sources. Such association is a precursor to 3D surface digitization where each laser plane develops a 3D surface. In this work, the detected laser line segments are modeled as a weighted directed acyclic graph (wDAG), in which the graph nodes represent the line segments and their maximum path hopes from any source node represents the laser source label. However, due to noise and occlusions, the structure of the wDAG will normally vary from that expected in the ideal cases. Furthermore, it is essential to robustness to incorporate the association information of the previous frames in the new frame associations. The proposed work objectively modifies the current frame’s wDAG under these constraints so the resulting wDAG becomes in an appropriate structure for the labeling of the detected line segments. Branch and bound (BnB) optimization has been a good choice for global optimization of combinatorial problems. Rather than approaching the optimization as a pure (or full) combinatorial problem, the proposed prioritizations of the branch modification actions are nearly optimal. The high priority actions will very likely lead to an optimal modification. This aspect is exploited by the proposed sparse variant of BnB (SBnB) optimization. It significantly reduces the number of the branches to be considered, in addition to those contributed by the bounding part of the BnB optimization. The proposed approach was tested on numerous monocular videos and plausible association results were efficiently achieved.

Parallel scheduling of multiple real-time applications onto heterogeneous processors is needed in the emerging embedded systems such as self-driving cars, smart cameras, and smartphones. Assuming that an embedded application is specified as a synchronous dataflow (SDF) graph or its extension, we propose a novel parallel scheduling methodology based on an evolutionary algorithm where the mapping of tasks onto processors is evolved to optimize a given objective function in an iterative fashion. In each iteration, we use an existing worst-case response time (WCRT) analysis tool to check if all applications satisfy their real-time requirements by translating each SDF graph into a directed acyclic graph (DAG) that is assumed in the WCRT analysis tool. Since the WCRT analysis must be performed in each iteration of evolution, we propose a clustering technique to reduce drastically the analysis time that depends on the number of nodes and their dependency. We formally prove that the proposed clustering technique does not change the estimated WCRT of each application. The effectiveness of the proposed scheduling methodology with the clustering technique is verified with extensive experiments using real-life benchmarks, randomly generated graphs, and the comparison with the existing technique.

Graph-based learning model has a wide range of applications in machine learning and computer vision. The key issue of the graph-based applications is to construct an informative graph to effectively represent data correlations. In practice, real-world data is usually contaminated by complex noise beyond Gaussian noise and sparse noise, which degrades learning performance dramatically. To construct a robust graph that represents real-world data distribution well, we propose a novel graph construction method. The proposed method is designed to be robust to complex noise beyond sparse or Gaussian noise by combining different previously given metrics of measuring data errors, which robustly and correctly generates a set of edges. To demonstrate the superiority of the proposed method, we employed the clustering and classification tasks on the real-world datasets. The experimental results demonstrate that the proposed method is superior to the existing methods in both accuracy and robustness to complex noise.

Graph embedding aims to represent node structural as well as attribute information into a low-dimensional vector space so that some downstream application tasks such as node classification, link prediction, community detection, and recommendation can be easily performed by using simple machine learning algorithms. The graph convolutional network is a neural network framework for machine learning on graphs. Because of its powerful ability to model graph data, it is currently the best choice for graph embedding. However, most existing graph convolutional network-based embedding algorithms not only ignore the data distribution of the latent codes but also lose the high-order proximity between nodes in a graph, leading to inferior embedding. To mitigate this problem, we investigate how to enforce latent codes to match a prior distribution, and we introduce random walk to preserve high-order proximity in a graph. In this paper, we propose a novel graph embedding framework, Adversarial and Random Walk Regularized Graph Embedding (ARWR-GE), which jointly preserves structural and attribute information. ARWR-GE adopts an adversarial training scheme to enforce the latent codes to match a prior distribution, and by employing the skip-gram model, nodes in a random walk sequence are closer in the latent space. We evaluate our proposed framework by using three real-world datasets on link prediction, graph clustering, and visualization tasks. The results demonstrate that our framework achieves better performance than state-of-the-art graph embedding algorithms.

In skeleton-based action recognition, the approaches based on graph convolutional networks(GCN) have achieved remarkable performance by modeling spatial-temporal graphs to explore the physical dependencies between body joints. However, these methods mostly apply hierarchical GCNs to aggregate wider-range neighborhood information, which makes joint features be weakened during long diffusion. In this paper, we design a multi-scale mixed dense graph convolutional network (MMDGCN) to overcome both shortcomings. We propose a dense graph convolution operation to enhance the local context information of joints, and then the spatial and temporal attention modules with a larger receptive field are introduced to help the model strengthen the discriminative features to adaptively refine the intermediate feature maps. We also design a multi-scale mixed temporal convolution module, which provides a flexible temporal graph through the combination of different scale convolution kernels. Extensive experiments on the three real-world datasets (NTU-RGB+D, NTU-RGB+D120 and Kinetics) demonstrate that the performance of the proposed MMDGCN in skeleton-based action recognition.

Intra-camera supervision (ICS) for person re-identification (Re-ID) assumes that identity labels are independently annotated within each camera view and no inter-camera identity association is labeled. It is a new setting proposed recently to reduce the burden of annotation while expect to maintain desirable Re-ID performance. However, the lack of inter-camera labels makes the ICS Re-ID problem much more challenging than the fully supervised counterpart. By investigating the characteristics of ICS, this article proposes a graph-induced contrastive learning (GCL) approach to address this issue. More specifically, we first formulate the cross-camera ID association task as a graph partitioning problem subjected to ICS-specific constraints and design a greedy agglomeration algorithm to solve it. Then, we propose a graph-induced contrastive loss that unifies both intra- and inter-camera learning into a contrastive learning framework to learn a Re-ID model. The cross-camera ID association step and the Re-ID model contrastive learning step are alternatively iterated, by which we progressively obtain a highly discriminative Re-ID model. Extensive experiments on three large-scale datasets show that our approach outperforms all previous ICS works. Especially, it gains 15.7% Rank-1 and 14.3% mAP improvements on the challenging MSMT17 dataset. Moreover, our approach performs even comparable to state-of-the-art fully supervised methods on all of the three datasets.

A significant task of automatic diagnosis for radiology imaging, especially for chest X-rays, is to identify disease types, which can be viewed as a multi-label learning problem. Prior state-of-the-art approaches adopted the graph convolutional network to model the correlations among disease labels. However, the utilization of medical reports paired with radiology images is neglected in such approaches. Hence, at least two novel improvements are proposed in this paper. First, disease label embeddings are pre-trained on the total radiology reports, and these semantic features along with encoded X-ray features are fused in a transformer encoder for graph initialization. Second, to expand the representation ability of the graph, extra medical terms from radiology reports are mined and added to the graph model as auxiliary nodes without changing the size of the output space. Experiments conducted on two public chest-X-ray datasets demonstrate the outstanding performance over compared models and the advantages of the proposed improvements.

For finding the defect inside of rails, we usually use the ultrasound to inspect the rails and obtain point clouds data of the returned ultrasonic signals after preprocessing. The quantity of points in each point cloud is not fixed, and the point clouds are disordered and unstructured. The points have heterogeneous attributes which include not only continuous coordinates but discrete frequencies of the ultrasound. For classifying the special point clouds, we propose a network architecture, named Multi-Branch Global Graph Convolution Network. For better utilizing the features of heterogeneous attributes of each point, we introduce the point channel-wise attention mechanism to weight the attributes. The backbone of the network is a multi-branch learn network based on the global graph convolution. By constructing the global graph and applying the graph convolution, the network possesses the ability to learn the local relative position information and global semantic information. Experimental results show the effectiveness of each component of the network and the state-of-the-art performance of the proposed network achieves on the classification task of defect recognition.

Existing popular methods for semi-supervised node classification with high-order convolution improve the learning ability of graph convolutional networks (GCNs) by capturing the feature information from high-order neighborhoods. However, these methods with high-order convolution usually require many parameters and high computational complexity. To address these limitations, we propose HCNP, a new higher-order GCN for semi-supervised node learning tasks, which can simultaneously aggregate information of various neighborhoods by constructing high-order convolution. In HCNP, we reduce the number of parameters using a weight sharing mechanism and combine the neighborhood information via multi-scale neighborhood pooling. Further, HCNP does not require a large number of hidden units, and it fits a few parameters and exhibits low complexity. We show that HCNP matches GCNs in terms of complexity and parameters. Comprehensive evaluations on publication citation datasets (Citeseer, Pubmed, and Cora) demonstrate that the proposed methods outperform MixHop in most cases while maintaining lower complexity and fewer parameters and achieve state-of-the-art performance in terms of accuracy and parameters compared to other baselines.

This paper introduces evolutionary game theory as an analysis tool to study the evolution of the cooperative behavior of nodes in the opportunistic network. At the same time, considering the spatial distribution of nodes in the opportunistic network and the corresponding cooperative interaction relationship, the evolutionary graph theory is further applied to study the cooperative behavior in the context of considering the network space. In the collaboration model of this article, the vertices in the evolutionary graph theory are used to represent the nodes in the network, and the edges in the graph represent that the node pairs are within the communication range of each other. The node will play a game with neighboring nodes according to the payout matrix. Through repeated games between nodes, the network reaches a relatively stable final state, and the distribution of nodes with different behaviors in the final state of the network is checked. By adjusting the network parameters, we can analyze the distribution and evolution of cooperation in the network, and study the stability of the network. We realize the credible collaborative evolutionary game model by building a comprehensive simulation platform and verify its effectiveness and scalability. We design and implement an evolutionary game model independent of routing protocols, so that the model can be quickly deployed in traditional routing protocols. Subsequently, simulations verified that the reputation model played a role in guiding nodes to actively participate in network collaboration during the collaborative evolution process. And the greater the strength of the reputation incentive mechanism, the better the effect of the evolutionary model. The evolution model is deployed in multiple routing protocols, verifying that the model can play a role in different routing protocols. In addition, the evolution model can also play a role in different network environments and has good scalability.

Graph-based methods have been widely used by the document image analysis and recognition community, as the different objects and the content in document images is best represented by this powerful structural representation. Designing of novel computation tools for processing these graph-based structural representations has always remained a hot topic of research. Recently, Graph Neural Network (GNN) have been used for solving different problems in the domain of document image analysis and recognition. In this article we take forward the state of the art by presenting a new approach to gather the symbolic and numeric information from the nodes and edges of a graph. We use this information to learn a Graph Neural Network (GNN). The experimentation on the recognition of handwritten letters and graphical symbols shows that the proposed approach is an interesting contribution to the growing set of GNN-based methods for document image analysis and recognition.

We show that precise knowledge of epidemic transmission parameters is not required to build an informative model of the spread of disease. We propose a detailed model of the topology of the contact network under various external control regimes and demonstrate that this is sufficient to capture the salient dynamical characteristics and to inform decisions. Contact between individuals in the community is characterised by a contact graph, the structure of that contact graph is selected to mimic community control measures. Our model of city-level transmission of an infectious agent (SEIR model) characterises spread via a (a) scale-free contact network (no control); (b) a random graph (elimination of mass gatherings); and (c) small world lattice (partial to full lockdown -- "social" distancing). This model exhibits good qualitative agreement between simulation and data from the 2020 pandemic spread of coronavirus. Estimates of the relevant rate parameters of the SEIR model are obtained and we demonstrate the robustness of our model predictions under uncertainty of those estimates. The social context and utility of this work is identified, contributing to a highly effective pandemic response in Western Australia., Revised. 14 pages, 7 figures. Including impact discussion and recovery predictions. in press in IEEE Access

Human skeleton contains intuitive information of actions and has high robustness in dynamic environment. Therefore, it has been widely studied in action recognition tasks. Most existing methods of skeleton recognition are based on graph convolutional networks (GCNs), which extract the topological structure of graphs to describe the dependencies between joints. However, the GCNs pay excessive attention to skeleton structure and neglect the feature information of skeleton joints. Accordingly, how to fuse the feature of both skeleton structure and joints is a problem to be solved. In addition, non-linear temporal convolutional network (TCN), which has more robustness and learning capability, is rarely investigated in existing methods. With the comprehensive consideration of the dependence between structure and feature on graphs, we propose a novel structure-feature fusion adaptive GCN (SFAGCN) for skeleton based action recognition. The topological structure of the skeleton graph and the feature of the joints can be fused by the decoupled spatiotemporal correlation in our model effectively. The relevance of spatiotemporal data can be preserved well by the fusion strategy, with data integrity ensured. Moreover, Gated TCN is used to extract temporal feature, improving the network performance further. We choose two-stream adaptive GCNs and shift-GCN as the baseline. To demonstrate the effectiveness of our methods, extensive experiments are implemented on the three large-scale datasets, namely, NTU-RGBD 60, NTU-RGBD 120 and Kinetics-Skeleton 400. The accuracy of top-1 on above datasets are improved by more than 0.6% on average, where the performance of SFAGCN exceeds the state-of-the-art methods.

In numerous applications the rapid increase of data size with time makes classical clustering algorithms too slow because of the high computational cost. In the present contribution, a novel method for online spectral clustering algorithm is introduced, which can be applied for simulation of structure and properties of titanate in chemical engineering. The proposed algorithm uses the recent results in the emerging field of graph signal processing. It avoids the costly computation of the eigenvectors by filtering random signals on the graph, and only a few derivative operations are needed to update the clustering result when a new data arrives. Initial simulations are presented using both simulated and real data sets, illustrating the relevance of the proposed algorithm.

In recent years, Graph Convolutional Networks (GCNs) have been increasingly and widely used in graph data representation and semi-supervised learning. GCNs can reveal and dig deep into irregular data with spatial topological structure. However, in the task of node classification, most models will be over-smoothing (indistinguishable representations of nodes in different classes) after stacking multilayer GCNs. To alleviate the issue of over-smoothing, we propose a novel Discriminable Squeeze and Excitation Graph Convolutional Network (D-SEGCN) based on the attention mechanism of features for semi-supervised node classification. In the proposed D-SEGCN model, Squeeze and Excitation (SE) module is fused to the Graph Convolutional Networks to form SEGCN, which can selectively enhance significant features and realize the adaptive calibration of feature dimensions, thus enhancing node features. Then, the feature representation is obtained through a hierarchical structure, it is put into a discriminable capsule layer which can judge the similarities between the node features to obtain new feature. Finally, the feature representation and new feature are fused to obtain the final feature to strengthen the discriminability of nodes. Furthermore, we demonstrate that D-SEGCN significantly outperforms the state-of-the-art methods on three citation networks datasets.

Next-hop data-ferrying is a data-driven approach to ferrying data between graph components of a disconnected network. In contrast to pre-planned routing methods like the Traveling Salesman tour, only the next-hop is planned, allowing the ferry to be reactive to the data flows within the network. When multiple ferries are used, explicit coordination between them is difficult due to the data-driven approach of the algorithm, and the distributed and disconnected nature of the problem. However, coordinated behavior can still be produced through swarm intelligent means and while this is a useful pragmatic result, mathematical models describing system properties are difficult to develop. This generally results in algorithm correctness being demonstrated through simulation. In this article, we describe the development of a mathematical model of such a next-hop ferrying swarm. We show that the closed-form expression for user data latency is a close match to simulation results for Random Geometric Graphs under the assumption of a zero expected degree. In non-zero degree graphs, a near constant offset between the model and simulation is observed.