Graph Structure Features Research Articles

Analysis of Knowledge Graph Path Reasoning Based on Variational Reasoning

Knowledge graph (KG) reasoning improves the perception ability of graph structure features, improving model accuracy and enhancing model learning and reasoning capabilities. This paper proposes a new GraphDIVA model based on the variational reasoning divergent autoencoder (DIVA) model. The network structures and calculation processes of the models are analyzed. The GraphSAGE algorithm is introduced into the path reasoning module to solve the inability of the original model to perceive the features of the graph structure, which leads to a decline in the accuracy rate. Hence, GraphDIVA can achieve a higher accuracy rate with fewer learning iterations. The experiments show the efficiency and effectiveness of our model and proves that our method has a better effect on the accuracy rate and training difficulty than the baseline model on the FB15k-237 and NELL-995 benchmark datasets.

SybilHunter: Hybrid graph-based sybil detection by aggregating user behaviors

Nowadays, online social networks (OSNs) support users in sharing their information and providing valuable data for various applications, e.g., rating and recommendation systems, search engine systems, etc. In such online social network-based applications, information quality is essential. However, the easy-to-use interactive user interfaces and free publication mechanisms facilitate malicious users to launch Sybil Attacks. They create fake identities, pollute user-generated content, and cause server information quality problems, such as buzz, rumor, spam, etc. Most existing graph-based sybil detection approaches only consider the static graph structure features and heavily rely on the prior knowledge of labeled nodes. “Limited-attack-edge” assumption, the critical assumption of these approaches has been proved invalid in many scenarios.In this paper, we propose SybilHunter, a hybrid graph-based sybil detection approach by aggregating user social behavior patterns. Our approach refines the OSN structure, quantifies nodes’ similarity according to the dynamic user behavior features to evaluate user pairs’ trustworthiness and consistency. Then it constructs a weighted-strong-social (WSS) graph, based on which SybilHunter outputs sybil nodes. We simulate and evaluate our approach under a Weibo dataset. The AUC of SybilHunter achieves 0.945, which is significantly higher than typical graph-based sybil detection methods, such as SybilRank, SybilWalk, etc.

Perturb more, trap more: Understanding behaviors of graph neural networks

While graph neural networks (GNNs) have shown great potential in various graph-related tasks, their lack of transparency has hindered our understanding of how they arrive at their predictions. The fidelity to the local decision boundary of the original model, indicating how well the explainer fits the original model around the instance to be explained, is neglected by existing GNN explainers. In this paper, we first propose a novel post hoc framework based on local fidelity for any trained GNNs, called TraP2, which can generate a high-fidelity explanation. Considering that both the relevant graph structure and important features inside each node must be highlighted, a three-layer architecture in TraP2 is designed: i) the interpretation domain is defined by the Translation layer in advance; ii) the local predictive behaviors of the GNNs being explained are probed and monitored by the Perturbation layer, in which multiple perturbations for graph structure and feature level are conducted in the interpretation domain; and iii) highly faithful explanations are generated by fitting the local decision boundary of GNNs being explained through the Paraphrase layer. We evaluated TraP2 on several benchmark datasets under the four metrics of accuracy, area under receiver operating characteristic curve, fidelity, and contrastivity, and the results prove that it significantly outperforms state-of-the-art methods.

A Graph Convolution Network Based on Improved Density Clustering for Recommendation System

Recommendation systems have been widely used in various applications to solve information overload andimprove user experience. Traditional recommendation algorithms mainly used Euclidean data for calculationand abandoned the graph structure features in user and item data. Aiming at the problems in the current recommendation algorithms, this paper proposes an improved user density clustering method and extracts userfeatures through optimized graph neural network. Firstly, the improved density clustering method is used toform the clustering subgraph of users based on the influence value of users. Secondly, the user data and itemdata features of cluster subgraph are extracted by graph convolution network. Finally, the features of clustersubgraphs are processed by global graph convolution network and the recommendation results are generatedaccording to the global graph features. This model not only improves the efficiency of decomposing large graphinto small graph through the improved user density clustering algorithm, but also extracts the features of usergroups through graph convolution neural network to improve the recommendation effect. The experiment alsoproves the validity of this model.

Graph Enhanced Neural Interaction Model for recommendation

Since user-item interactions in recommender systems can be naturally modeled as a bipartite graph, recent studies have started to incorporate graph neural networks (GNNs) to learn user and item representations. However, existing GNN-based models for recommendation usually emphasize the graph structure but neglect the rich node (i.e., users and items) features or linearly integrate the node features without considering the interactions among these features, which leads to a suboptimal recommendation model. Considering the features (e.g., categorical features) are often discrete, sparse, and high-dimensional, how to jointly utilize the graph structure and node features while considering the feature interactions is a major challenge. In this paper, we propose Graph Enhanced Neural Interaction Model (GENIM), a novel graph recommendation model consisting of three parts: (1) graph convolution layers that recursively propagate the encoded node features on the user-item bipartite graph; (2) the neural feature interaction layer that learns node feature interactions, which contains rich signals for predictive analytics; (3) an optional hashing-based embedding layer that is used to reduce the model size. Extensive experiments conducted on two real-world datasets show that our model outperforms other state-of-the-art solutions.

Detection of DDoS Attack within Industrial IoT Devices Based on Clustering and Graph Structure Features

Network available and accessible is of great importance to the Internet of things (IoT) devices. In this study, a novel machine learning method is presented to predict the occurrence of distributed denial-of-service (DDoS) attacks. Firstly, a structure of edges and vertices within graph theory is created to simultaneously extract traffic data characteristics. Eight characteristics of traffic data are selected as input variables. Secondly, the principal component analysis (PCA) model is adopted to extract DDoS and normal communication features further. Then, DDoSs are detected by fuzzy C-means (FCM) clustering with these features. In the case study, 2000 traffic data in dataset CICIDS-2017 are used to verify the practicability of this method. The results of recall, false positive, true positive, true negative, and false negative are 100.00%, 1.05%, 68.95%, 0.00%, and 30.00%. Compared with other methods, the results demonstrate that the detecting reliability is improved, and the method has a good effect on the detection of DDoS attacks.

LWP-WL: Link weight prediction based on CNNs and the Weisfeiler–Lehman algorithm

We present a new technique for link weight prediction, the Link Weight Prediction Weisfeiler–Lehman (LWP-WL) method that learns from graph structure features and link relationship patterns. Inspired by the Weisfeiler–Lehman Neural Machine, LWP-WL extracts an enclosing subgraph for the target link and applies a graph labelling algorithm for weighted graphs to provide an ordered subgraph adjacency matrix into a neural network. The neural network contains a Convolutional Neural Network in the first layer that applies special filters adapted to the input graph representation. An extensive evaluation is provided that demonstrates an improvement over the state-of-the-art methods in several weighted graphs. Furthermore, we conduct an ablation study to show how adding different features to our approach improves our technique’s performance. Finally, we also perform a study on the complexity and scalability of our algorithm. Unlike other approaches, LWP-WL does not rely on a specific graph heuristic and can perform well in different kinds of graphs.

Topological Graph Convolutional Network Based on Complex Network Characteristics

Graph convolutional neural networks have received a lot of attention in various tasks dealing with graph data by aggregating information from neighboring nodes and passing node information. Many recent studies have looked at the impact of topological features on node classification tasks by altering aggregation based on degree values or incorporating topological data analysis into graph convolutional neural networks; however, graph data itself has many topological characteristics in complex networks. In most circumstances, the graph’s topological characteristics reveal the nodes’ similarity and facilitate the node classification task. This paper proposes a topological structure feature extraction method based on the concept of complex topological characteristics, which can obtain deeper topological features in the graph structure and use node features to obtain feature information that is more important for classification from both feature spaces. Evidence from experimental studies has established that the topological structure obtained by the method in this paper can be used as input to the GCN,and good results can be achieved on the classification task even without any external information on the nodes. In the graph dataset of connected topologies, the method exhibits a very large increase in accuracy and macro F1-score when compared to the state-of-the-art baseline model after mixing the node features.

Spectral embedding network for attributed graph clustering

Attributed graph clustering aims to discover node groups by utilizing both graph structure and node features. Recent studies mostly adopt graph neural networks to learn node embeddings, then apply traditional clustering methods to obtain clusters. However, they usually suffer from the following issues: (1) they adopt original graph structure which is unfavorable for clustering due to its noise and sparsity problems; (2) they mainly utilize non-clustering driven losses that cannot well capture the global cluster structure, thus the learned embeddings are not sufficient for the downstream clustering task. In this paper, we propose a spectral embedding network for attributed graph clustering (SENet), which improves graph structure by leveraging the information of shared neighbors, and learns node embeddings with the help of a spectral clustering loss. By combining the original graph structure and shared neighbor based similarity, both the first-order and second-order proximities are encoded into the improved graph structure, thus alleviating the noise and sparsity issues. To make the spectral loss well adapt to attributed graphs, we integrate both structure and feature information into kernel matrix via a higher-order graph convolution. Experiments on benchmark attributed graphs show that SENet achieves superior performance over state-of-the-art methods.

Semi-AttentionAE: An Integrated Model for Graph Representation Learning

Graph embedding learns low-dimensional vector representations which capture and preserve information in original graphs. Common shallow neural networks and deep autoencoder only use adjacency matrix as input, and usually ignore node attributes and features. Shallow graph neural networks cannot spread the node characteristic information on a large scale. Many deep models suffer the problem of over-smoothing. Therefore, these methods can't fully incorporate network information. In this paper, we propose a novel Semi-AttentionAE model to fully utilize node features, node labels, and network structure. More specifically, we integrate a supervised information extraction graph attention network to capture both node features and network structure, with an unsupervised feature extraction autoencoder to reduce dimension while preserving structure information. Finally, ensemble learning is introduced to jointly train the combined model to obtain final embedding. We conduct the node classification and visualization experiments on four real-world datasets, including two citation networks, one co-occurrence network, and one commodity network. The results suggest that the proposed Semi-AttentionAE model is capable of embedding both graph structure and node features. The integrated model has successfully exceeded or matched performance across four well-established baselines.

A novel framework of graph Bayesian optimization and its applications to real-world network analysis

Network structure optimization is a fundamental task of many expert and intelligent systems, such as the intelligent tools for chemical molecular discovery and expert systems for road network design. However, traditional model-free methods suffer from the expensive computational cost of evaluating networks; almost all the research on Bayesian optimization is aimed at optimizing the objective functions with vectorial inputs, e.g., the hyper-parameters in any expert systems. This work focuses on applying Bayesian optimization to optimize network structure with graph-structured inputs, and presents a flexible framework, denoted as graph Bayesian optimization (GBO), to handle arbitrary graphs. By combining the proposed framework with graph kernels, it can take full advantage of implicit graph structural features to supplement explicit features guessed according to the experience, such as tags of nodes and any attributes of graphs. Simultaneously, the proposed framework can identify which features are more important during the optimization process. By collaboratively working with a down-stream decision tree, the GBO can not only find the optimum but also discover its knowledge represented by rules, which can further enhance its interpretability and assist expert decision-making. A novel problem of opening the gated residential areas is presented in this work, which can serve as one benchmark task of road network design. Intensive experiments conducted on three real-world applications, including robust network design, the most active node identification, and urban transportation network design, demonstrate its efficacy and potential applications.

Seven-Layer Model in Complex Networks Link Prediction: A Survey.

Link prediction is the most basic and essential problem in complex networks. This study analyzes the observed topological, time, attributive, label, weight, directional, and symbolic features and auxiliary information to find the lack of connection and predict the future possible connection. For discussion and analysis of the evolution of the network, the network model is of great significance. In the past two decades, link prediction has attracted extensive attention from experts in various fields, who have published numerous high-level papers, but few combine interdisciplinary characteristics. This survey analyzes and discusses the existing link prediction methods. The idea of stratification is introduced into the classification system of link prediction for the first time and proposes the design idea of a seven-layer model, namely the network, metadata, feature classification, selection input, processing, selection, and output layers. Among them, the processing layer divides link prediction methods into similarity-based, probabilistic, likelihood, supervised learning, semi-supervised learning, unsupervised learning, and reinforcement learning methods. The input features, evaluation metrics, complex analysis, experimental comparisons, relative merits, common dataset and open-source implementations for each link prediction method are then discussed in detail. Through analysis and comparison, we found that the link prediction method based on graph structure features has better prediction performance. Finally, the future development direction of link prediction in complex networks is discussed.

Anisotropic Graph Convolutional Network for Semi-Supervised Learning

Graph convolutional networks learn effective node embeddings that have proven to be useful in achieving high-accuracy prediction results in semi-supervised learning tasks, such as node classification. However, these networks suffer from the issue of over-smoothing and shrinking effect of the graph due in large part to the fact that they diffuse features across the edges of the graph using a linear Laplacian flow. This limitation is especially problematic for the task of node classification, where the goal is to predict the label associated with a graph node. To address this issue, we propose an anisotropic graph convolutional network for semi-supervised node classification by introducing a nonlinear function that captures informative features from nodes, while preventing oversmoothing. The proposed framework is largely motivated by the good performance of anisotropic diffusion in image and geometry processing, and learns nonlinear representations based on local graph structure and node features. The effectiveness of our approach is demonstrated on three citation networks and two image datasets, achieving better or comparable classification accuracy results compared to the standard baseline methods.

Attributed network representation learning via improved graph attention with robust negative sampling

Attributed network representation learning is to embed graphs in low dimensional vector space such that the embedded vectors follow the differences and similarities of the source graphs. To capture structural features and node attributes of attributed network, we propose a novel graph auto-encoder method which is stacked encoder-decoder layers based on graph attention with robust negative sampling. Here, minimize the negative log-likelihood, triplet distance, and weighted neighborhood attributes are proposed as the loss function. To alleviate the over-fitting on reconstruct graph structural features or node attributes, a trade off algorithm between reconstruction loss of node attributes and reconstruction loss of structural features is proposed. Furthermore, to alleviate the impact of random sampling, we propose additional constraints on negative sampling based on node degree. Experimental results on several benchmark datasets for transductive and inductive learning tasks show that the proposed model is competitive against well-known methods in node classification and link prediction.

Graph convolutional autoencoder model for the shape coding and cognition of buildings in maps

ABSTRACT The shape of a geospatial object is an important characteristic and a significant factor in spatial cognition. Existing shape representation methods for vector-structured objects in the map space are mainly based on geometric and statistical measures. Considering that shape is complicated and cognitively related, this study develops a learning strategy to combine multiple features extracted from its boundary and obtain a reasonable shape representation. Taking building data as example, this study first models the shape of a building using a graph structure and extracts multiple features for each vertex based on the local and regional structures. A graph convolutional autoencoder (GCAE) model comprising graph convolution and autoencoder architecture is proposed to analyze the modeled graph and realize shape coding through unsupervised learning. Experiments show that the GCAE model can produce a cognitively compliant shape coding, with the ability to distinguish different shapes. It outperforms existing methods in terms of similarity measurements. Furthermore, the shape coding is experimentally proven to be effective in representing the local and global characteristics of building shape in application scenarios such as shape retrieval and matching.

DDoS detection based on graph structure features and non‐negative matrix factorization

AbstractDistributed denial of service (DDoS) attacks are the most common and harmful attack in the field of network security, the purpose of this paper is to predict the occurrence of DDoS attacks quickly, accurately and effectively. In this paper, the characteristics of DDoS attacks are analyzed in detail, graph model is used to express the structural characteristics of traffic data between the victims and bots, at the same time, considering data characteristics of traffic itself, 11 characteristics are selected at last. The method of nonnegative matrix factorization (NMF) is adopted to mine the behavior pattern of DDoS attacks, and then the DDoS bots is detected by clustering according to the projection value of nodes on the behavior pattern. By extracting the attack behavior pattern, the attack behavior can be detected. Experimental results of real dataset verify the effectiveness and validity of the method.

A Multi-Scale Approach for Graph Link Prediction

Deep models can be made scale-invariant when trained with multi-scale information. Images can be easily made multi-scale, given their grid-like structures. Extending this to generic graphs poses major challenges. For example, in link prediction tasks, inputs are represented as graphs consisting of nodes and edges. Currently, the state-of-the-art model for link prediction uses supervised heuristic learning, which learns graph structure features centered on two target nodes. It then learns graph neural networks to predict the existence of links based on graph structure features. Thus, the performance of link prediction models highly depends on graph structure features. In this work, we propose a novel node aggregation method that can transform the enclosing subgraph into different scales and preserve the relationship between two target nodes for link prediction. A theory for analyzing the information loss during the re-scaling procedure is also provided. Graphs in different scales can provide scale-invariant information, which enables graph neural networks to learn invariant features and improve link prediction performance. Our experimental results on 14 datasets from different areas demonstrate that our proposed method outperforms the state-of-the-art methods by employing multi-scale graphs without additional parameters.

A Novel Time Constraint-Based Approach for Knowledge Graph Conflict Resolution

Knowledge graph conflict resolution is a method to solve the knowledge conflict problem in constructing knowledge graphs. The existing methods ignore the time attributes of facts and the dynamic changes of the relationships between entities in knowledge graphs, which is liable to cause high error rates in dynamic knowledge graph construction. In this article, we propose a knowledge graph conflict resolution method, knowledge graph evolution algorithm based on deep learning (Kgedl), which can resolve facts confliction with high precision by combing time attributes, semantic embedding representations, and graph structure features. Kgedl first trains the semantic embedding vector through the relationships between entities. Then, the path embedding vector is trained from the graph structures of knowledge graphs, and the time attributes of entities are combined with the semantic and path embedding vectors. Finally, Kgedl uses a recurrent neural network to make the inconsistent facts appear in the dynamic evolution of the knowledge graph consistent. A large number of experiments on real datasets show that Kgedl outperforms the state-of-the-art methods. Especially, Kgedl achieves 23% higher performance than the classical method numerical Probabilistic Soft Logic (nPSL).in the metric HITS@10. Also, extensive experiments verified that our proposal possess better robustness by adding noise data.

Cascade embedding model for knowledge graph inference and retrieval

Knowledge graphs are widely used in retrieval systems, question answering systems (QA), hypothesis generation systems, etc. Representation learning provides a way to mine knowledge graphs to detect missing relations; and translation-based embedding models are a popular form of representation model. Shortcomings of translation-based models however, limits their practicability as knowledge completion algorithms. The proposed model helps to address some of these shortcomings.The similarity between graph structural features of two entities was found to be correlated to the relations of those entities. This correlation can help to solve the problem caused by unbalanced relations and reciprocal relations. We used Node2vec, a graph embedding algorithm, to represent information related to an entity's graph structure, and we introduce a cascade model to incorporate graph embedding with knowledge embedding into a unified framework. The cascade model first refines feature representation in the first two stages (Local Optimization Stage), and then uses backward propagation to optimize parameters of all the stages (Global Optimization Stage). This helps to enhance the knowledge representation of existing translation-based algorithms by taking into account both semantic features and graph features and fusing them to extract more useful information. Besides, different cascade structures are designed to find the optimal solution to the problem of knowledge inference and retrieval.The proposed model was verified using three mainstream knowledge graphs: WIN18, FB15K and BioChem. Experimental results were validated using the hit@10 rate entity prediction task. The proposed model performed better than TransE, giving an average improvement of 2.7% on WN18, 2.3% on FB15k and 28% on BioChem. Improvements were particularly marked where there were problems with unbalanced relations and reciprocal relations. Furthermore, the stepwise-cascade structure is proved to be more effective and significantly outperforms other baselines.

Graph Embedding Matrix Sharing With Differential Privacy

Graph embedding maps a graph into low-dimensional vectors, i.e., embedding matrix, while preserving the graph structure, solving the high computation and space cost for graph analysis. Matrix factorization (MF) is an effective means to achieve graph embedding since maintaining the utility of the graph structure. The personalized graph structure features implied in the embedding matrix can identify the individual, which potentially breaches individual sensitive information in the original graph. Currently, protecting individual privacy without compromising the utility is the key to sharing the embedding matrix. Differential privacy is a gold standard for publishing sensitive information while protecting privacy. The existing methods on differentially private MF, however, cannot be directly incorporated onto MF-based graph embedding as they undergo either high global sensitivity or iterative noise error accumulation, potentially rendering poor utility of MF-based graph embedding. To address the deficiency, this study proposes PPGD, a differentially private perturbed gradient descent method for MF-based graph embedding matrix sharing. Specifically, a Lipschitz condition on the objective function of the MF and a gradient clipping strategy are devised for bounding global sensitivity. Along the way, a scalable solution to global sensitivity that is independent on the original dataset is proposed. Further, a composite noise added means in the gradient descent is designed to guarantee privacy while enhancing the utility. The theoretical analysis shows that PPGD can generate processed embedding matrix with the utility maximization while achieving (e, δ)-differential privacy. The experimental evaluations confirm the effectiveness and efficiency of PPGD.