Graph Structure Features Research Articles

USAGE : Uncertain flow graph and spatio-temporal graph convolutional network-based saturation attack detection method

With the development of Software-Defined Networking (SDN), its security problems have attached interests from academia. The saturation attack targeted at the SDN switch is one of the most destructive problems in SDN. The saturation attack can destroy SDN networks by depleting the computation and storage resources of SDN switches. Currently, graph neural networks have already been used for detecting saturation attacks in SDN. When constructing the graph based on K-Nearest Neighbors (KNN) algorithm, the number of edges in the graph is fixed. That may fail on accurately representing the relationship among different nodes. In addition, when using the Graph Convolutional Network (GCN) to detect saturation attack against SDN, only a partial graph structure and node features will be learned, causing the flow graph nodes information cannot be fully utilized. To overcome these issues, in this work, we propose USAGE, an uncertain flow graph and spatio-temporal graph neural network-based saturation attack detection method. More specifically, USAGE designs the CRAM algorithm and EGG model. The CRAM algorithm is proposed to generate an uncertain flow graph for each network traffic flow, which can represent the uncertain relationship among different nodes. EGG is a spatio-temporal graph convolutional network model, which is proposed to detect saturation attacks. The evaluation results demonstrate that USAGE can generate a better graph than KNN and achieve high detection performance. Meanwhile, USAGE outperforms other methods in terms of TPR, TNR, accuracy, precision, recall, F-score, confusion matrix, ROC curves, and PR curves.

A Knowledge-based and Data-driven Approach for Predicting Acute Kidney Injury in Patients with Heart Failure.

It has great potential to integrate medical knowledge and electronic health record data for diagnosis prediction. However, present studies only utilized information from knowledge graphs, omitting potentially significant global graph structural features. In this study, we proposed a knowledge and data integrating modeling approach to reconstruct patient electronic health record data with graph structure and use medical knowledge as internal information of patient data to build a risk prediction model for acute kidney injury in patients with heart failure based on graph neural networks. Experimental results based on the MIMIC III data showed that the method proposed was superior to other baseline models in predicting the risk of acute kidney injury in heart failure patients, with an accuracy of 0.725 and an F1 score of 0.755. This study provides a novel approach to the disease risk prediction models that integrates medical knowledge and data.

Air quality prediction by integrating mechanism model and machine learning model

AQP (Air Quality Prediction) is a very challenging project, and its core issue is how to solve the interaction and influence among meteorological, spatial and temporal factors. To address this central conundrum, we make full use of the characteristics of mechanism model and machine learning and propose a new AQP method based on DM_STGNN (Dynamic Multi-granularity Spatio-temporal Graph Neural Network). This method is the first time to use the air quality model HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory Model) to assist in building a dynamic spatio-temporal graph structure to learn the spatiotemporal relationship of pollutants. DM_STGNN is based on an elaborate encoder-decoder architecture. At the encoder, in order to better mine the spatial dependency, we built a multi-granularity graph structure, used meteorological, time and geographical features to establish node attributes, used well-known HYSPLIT model to dynamically establish the edges among nodes, and used LSTM (Long Short Term Memory) to learn the time-series relationship of pollutant concentrations. At the decoder, in order to better mine the temporal dependency, we built an attention mechanism based LSTM for decoding and AQP. Additionally, in order to efficiently learn the temporal patterns from very long-term historical time series and generate rich contextual information, an unsupervised pre-training model is used to enhance DM_STGNN. The proposed model makes full use of and fully considers the influence of meteorological, spatial and temporal factors, and integrates the advantages of mechanism model and machine learning. On a project-based dataset, we validate the effectiveness of the proposed model and examine its abilities of capturing both fine-grained and long-term influences in AQP. We also compare the proposed model with the state-of-the-art AQP methods on the dataset of Yangtze River Delta city group, the experimental results show the appealing performance of our model over competitive baselines.

Topological Pooling on Graphs

Graph neural networks (GNNs) have demonstrated a significant success in various graph learning tasks, from graph classification to anomaly detection. There recently has emerged a number of approaches adopting a graph pooling operation within GNNs, with a goal to preserve graph attributive and structural features during the graph representation learning. However, most existing graph pooling operations suffer from the limitations of relying on node-wise neighbor weighting and embedding, which leads to insufficient encoding of rich topological structures and node attributes exhibited by real-world networks. By invoking the machinery of persistent homology and the concept of landmarks, we propose a novel topological pooling layer and witness complex-based topological embedding mechanism that allow us to systematically integrate hidden topological information at both local and global levels. Specifically, we design new learnable local and global topological representations Wit-TopoPool which allow us to simultaneously extract rich discriminative topological information from graphs. Experiments on 11 diverse benchmark datasets against 18 baseline models in conjunction with graph classification tasks indicate that Wit-TopoPool significantly outperforms all competitors across all datasets.

Quaternion-based graph convolution network for recommendation

Graph Convolution Network (GCN) has been widely applied in recommender systems for its representation learning capability on user and item embeddings. However, GCN is vulnerable to noisy and incomplete graphs, which are common in real world, due to its recursive message propagation mechanism. In the literature, some work proposes to remove the feature transformation during message propagation, but making it unable to effectively capture the graph structural features. Moreover, they model users and items in the Euclidean space, which has been demonstrated to have high distortion when modeling complex graphs, further degrading the capability to capture the graph structural features and leading to sub-optimal performance. To this end, in this paper, we propose a simple yet effective Quaternion-based Graph Convolution Network (QGCN) recommendation model. In the proposed model, we utilize the hyper-complex Quaternion space to learn user and item representations and feature transformation to improve both performance and robustness. Specifically, we first embed all users and items into the Quaternion space. Then, we introduce the quaternion embedding propagation layers with quaternion feature transformation to perform message propagation. Finally, we combine the embeddings generated at each layer with the mean pooling strategy to obtain the final embeddings for recommendation. Extensive experiments on three public benchmark datasets demonstrate that our proposed QGCN model outperforms baseline methods by a large margin.

A Novel Link Prediction Method for Social Multiplex Networks Based on Deep Learning

Due to the great advances in information technology, an increasing number of social platforms have appeared. Friend recommendation is an important task in social media, but newly built social platforms have insufficient information to predict entity relationships. In this case, platforms with sufficient information can help newly built platforms. To address this challenge, a model of link prediction in social multiplex networks (LPSMN) is proposed in this work. Specifically, we first extract graph structure features, latent features and explicit features and then concatenate these features as link representations. Then, with the assistance of external information from a mature platform, an attention mechanism is employed to construct a multiplex and enhanced forecasting model. Additionally, we consider the problem of link prediction to be a binary classification problem. This method utilises three different kinds of features to improve link prediction performance. Finally, we use five synthetic networks with various degree distributions and two real-world social multiplex networks (Weibo–Douban and Facebook–Twitter) to build an experimental scenario for further assessment. The numerical results indicate that the proposed LPSMN model improves the prediction accuracy compared with several baseline methods. We also find that with the decline in network heterogeneity, the performance of LPSMN increases.

Multi-View Tensor Graph Neural Networks Through Reinforced Aggregation

Graph Neural Networks (GNNs) have yielded fruitful results in learning multi-view graph data. However, it is challenging for existing GNNs to capture the potential correlation information (PCI) among the graph structure features of multiple views. It is also challenging to adaptively identify valuable neighbors for node feature fusion in different views. To this end, we propose a novel <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</u> einforced <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">T</u> ensor <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">G</u> raph <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</u> eural <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">N</u> etwork (RTGNN) framework to more effectively perform multi-view graph representation learning through reinforcing inter- and intra-graph aggregation. Specifically, RTGNN first uses tensor decomposition to extract the graph structure features (GSFs) of each view in the common feature space. These GSFs contain the PCI of multiple views and alleviate fusion conflicts that may be caused by differences between view feature spaces in cross-view feature fusion. Since fusing the features of all neighbor nodes may harm the features of the center node, we filter the irrelevant neighbors to improve the performance of intra-graph aggregation in each view. Concretely, a reinforcement learning (RL)-guided scheme is developed to automatically calculate the optimal filtering threshold for each view, avoiding tedious manual updates and infeasible back propagation updates. Experimental results and analysis on five datasets show that RTGNN surpasses the best multi-view graph representation baselines and achieves the maximum 14.26% performance improvement in terms of F1. The code link is <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/RingBDStack/RTGNN</uri> .

EEG-Based Emotional Video Classification via Learning Connectivity Structure

Electroencephalography (EEG) is a useful way to implicitly monitor the users perceptual state during multimedia consumption. One of the primary challenges for the practical use of EEG-based monitoring is to achieve a satisfactory level of accuracy in EEG classification. Connectivity between different brain regions is an important property for the classification of EEG. However, how to define the connectivity structure for a given task is still an open problem, because there is no ground truth about how the connectivity structure should be in order to maximize the classification performance. In this paper, we propose an end-to-end neural network model for EEG-based emotional video classification, which can extract an appropriate multi-layer graph structure and signal features directly from a set of raw EEG signals and perform classification using them. Experimental results demonstrate that our method yields improved performance in comparison to the existing approaches where manually defined connectivity structures and signal features are used. Furthermore, we show that the graph structure extraction process is reliable in terms of consistency, and the learned graph structures make much sense in the viewpoint of emotional perception occurring in the brain.

FSopt_k: Finding the Optimal Anonymization Level for a Social Network Graph

k-degree anonymity is known as one of the best models for anonymizing social network graphs. Although recent works have tried to address the privacy challenges of social network graphs, privacy levels are considered to be independent of the features of the graph degree sequence. In other words, the optimal value of k is not considered for the graph, leading to increasing information loss. Additionally, the graph may not need a high privacy level. In addition, determining the optimal value of k for the graph in advance is a big problem for the data owner. Therefore, in this paper, we present a technique named FSopt_k that is able to find the optimal value of k for each social network graph. This algorithm uses an efficient technique to partition the graph nodes to choose the best k value. It considers the graph structure features to determine the best privacy level. In this way, there will be a balance between privacy and loss in the anonymized graph. Furthermore, information loss will be as low as possible. The evaluation results depict that this algorithm can find the optimal value of k in a short time as well as preserve the graph’s utility.

Multi-view representation model based on graph autoencoder

Graph representation learning is a hot topic in non-Euclidean data in various domains, such as social networks, biological networks, etc. When some data labels are missing, graph autoencoder and graph variational autoencoder can perform outstanding abilities on node clustering or link prediction tasks. However, most existing graph representation learning ignores data's multi-modal features and takes the node features and graph structure features as one view. Besides, most graph autoencoders only reconstruct the node feature matrix or adjacency matrix, which does not fully use the hidden representation information. In this paper, we propose a multi-view representation model based on graph autoencoder, which can employ the global structure topology, latent local topology, and feature relative information. Meanwhile, we add another decoder to reconstruct the node feature matrix as an auxiliary task. In this way, the proposed framework can utilize the learned representation sufficiently. We validate the effectiveness of our framework on four datasets, and the experimental results demonstrate superior performance compared with other advanced frameworks.

Detecting Phishing Accounts on Ethereum Based on Transaction Records and EGAT

In recent years, the losses caused by scams on Ethereum have reached a level that cannot be ignored. As one of the most rampant crimes, phishing scams have caused a huge economic loss to blockchain platforms and users. Under these circumstances, to address the threat to the financial security of blockchain, an Edge Aggregated Graph Attention Network (EGAT) based on the static subgraph representation of the transaction network is proposed. This study intends to detect Ethereum phishing accounts through the classification of transaction network subgraphs with the following procedures. Firstly, the accounts are used as nodes and the flow of transaction funds is used as directed edges to construct the transaction network graph. Secondly, the transaction record data of phishing accounts in the publicly available Ethereum are analyzed and statistical features of Value, Gas, and Timestamp values are manually constructed as node and edge features of the graph. Finally, the features are extracted and classified using the EGAT network. According to the experimental results, the Recall of the proposed method from the article is 99.3% on the dataset of phishing accounts. As demonstrated, the EGAT is more efficient and accurate compared with Graph2Vec and DeepWalk, and the graph structure features can express semantics better than manual features and simple transaction networks, which effectively improves the performance of phishing account detection.

MLSFF: Multi-level structural features fusion for multi-modal knowledge graph completion.

With the rise of multi-modal methods, multi-modal knowledge graphs have become a better choice for storing human knowledge. However, knowledge graphs often suffer from the problem of incompleteness due to the infinite and constantly updating nature of knowledge, and thus the task of knowledge graph completion has been proposed. Existing multi-modal knowledge graph completion methods mostly rely on either embedding-based representations or graph neural networks, and there is still room for improvement in terms of interpretability and the ability to handle multi-hop tasks. Therefore, we propose a new method for multi-modal knowledge graph completion. Our method aims to learn multi-level graph structural features to fully explore hidden relationships within the knowledge graph and to improve reasoning accuracy. Specifically, we first use a Transformer architecture to separately learn about data representations for both the image and text modalities. Then, with the help of multimodal gating units, we filter out irrelevant information and perform feature fusion to obtain a unified encoding of knowledge representations. Furthermore, we extract multi-level path features using a width-adjustable sliding window and learn about structural feature information in the knowledge graph using graph convolutional operations. Finally, we use a scoring function to evaluate the probability of the truthfulness of encoded triplets and to complete the prediction task. To demonstrate the effectiveness of the model, we conduct experiments on two publicly available datasets, FB15K-237-IMG and WN18-IMG, and achieve improvements of 1.8 and 0.7%, respectively, in the Hits@1 metric.

MultiBiSage

Graph Convolutional Networks (GCN) can efficiently integrate graph structure and node features to learn high-quality node embeddings. At Pinterest, we have developed and deployed PinSage, a data-efficient GCN that learns pin embeddings from the Pin-Board graph. Pinterest relies heavily on PinSage which in turn only leverages the Pin-Board graph. However, there exist several entities at Pinterest and heterogeneous interactions among these entities. These diverse entities and interactions provide important signal for recommendations and modeling. In this work, we show that training deep learning models on graphs that captures these diverse interactions can result in learning higher-quality pin embeddings than training PinSage on only the Pin-Board graph. However, building a large-scale heterogeneous graph engine that can process the entire Pinterest size data has not yet been done. In this work, we present a clever and effective solution where we break the heterogeneous graph into multiple disjoint bipartite graphs and then develop novel data-efficient MultiBiSage model that combines the signals from them. MultiBiSage can capture the graph structure of multiple bipartite graphs to learn high-quality pin embeddings. The benefit of our approach is that individual bipartite graphs can be processed with minimal changes to Pinterest's current infrastructure, while being able to combine information from all the graphs while achieving high performance. We train MultiBiSage on six bipartite graphs including our Pin-Board graph and show that it significantly outperforms the deployed latest version of PinSage on multiple user engagement metrics. We also perform experiments on two public datasets to show that MultiBiSage is generalizable and can be applied to datasets outside of Pinterest.

A Semantic Mapping Method of Relation Representation Enhancement for Few-Shot Knowledge Graph Completion

Few-shot knowledge graph completion (FKGC) tasks involve determining the authenticity of triple candidates using a small number of reference triples with a given relation. Intuitively, the expression of relation features contributes to the close correlation among the triples with the same relation. Therefore, the relation features are not comprehensive enough, leading to the fact that the triples cannot learn sufficient association information for the FKGC task. In this paper, an enhanced relation semantic representation model is constructed for associative reference triples from both aspects of external structure and internal semantics. On the one hand, as the structure around the triple is helpful to understand the relation semantics implicated in the triple, a graph convolution network with attention and relation features is proposed to obtain the graph structure features. Furthermore, the local structure information of triples can be used to learn the deep relation semantics. On the other hand, entity information can enhance the perception of the relation semantics in the triple. Afterward, in order to associate the triples with the same relation by enhanced relation semantics, a semantic mapping method is proposed, which uses shared merged variables to map the relation, entity, and graph structure features into the same embedding space. Finally, a prototype network based on attention convolution is established to extract the relation prototype representation, and then classify query triples to achieve the purpose of completing the knowledge graph. Through experimental verification, the proposed model achieves excellent performance on two datasets commonly used in the FKGC.

Sensor Placement for Leak Localization in Water Distribution Networks Based on Graph Convolutional Network

Nowadays, the research on optimal sensor placement (OSP) is very active in the field of leak detection and localization in the water distribution network (WDN). Considering that most of the WDNs are irregular networks, this article proposes a new method for OSP applied to irregular network system conditions. From the initial set of a limited number of sensors placed by node data features, the proposed method makes a completely new change to the limited number of sensor layouts. The solution is based on a graph convolutional network (GCN) algorithm, which combines node features and graph structure features. In order to verify the effectiveness of the algorithm, the nodes selected by the algorithm in this article are used to locate the leakage of the pipeline network. By simulating real leakage scenarios, the method is applied to a small WDN (Anytown) and a large WDN (Net3). The results show that the method proposed in this article outperforms other comparison algorithms to a certain extent.

A graph structure feature-based framework for the pattern recognition of the operational states of integrated energy systems

The recognition of operational state patterns is crucial for the safe, reliable and profitable operational management of integrated energy systems (IES). However, considering that the monitored operational data are often not labelled, supervised machine learning classification methods cannot be directly applied. In this study, based on the idea of graph theory, a framework of interpretable time series pattern recognition methods based on graph structural features is proposed. Firstly, the time series data is segmented into local subsequences by a shapelet-based time series classification method. Secondly, time-graphlets are constructed by means of a visibility graph algorithm and an original graph structure feature-based representation method is proposed. Then, pattern recognition of the different operational states of IES is performed in the form of multidimensional feature vectors. Thirdly, an association rule model is designed to analyze the supply and demand fluctuations, price fluctuations and couplings between multiple energy subsystems, based on which the pattern recognition results are annotated to match the operational characteristics of the IES. Fourthly, a shapelet-based state transfer network graph is constructed, and the complex network centrality metrics are used to analyze the operating characteristics and key operating states of the system over time. The new method is applied for pattern recognition of operational states in a Spanish IES. The results show that the proposed method can increase the interpretability of the recognition results and provide explanatory labels for the operational data. Furthermore, the vulnerable operational states of the system can be revealed, which allows managers to monitor the operational state of the system and take early defensive measures against the more threatening operational patterns. This suggests that the methodological framework presented in this paper is a promising approach to pattern recognition that can provide interpretable annotations for unsupervised operational data of IES and their safe management.

Dynamic scheduling of tasks in cloud manufacturing with multi-agent reinforcement learning

Cloud manufacturing provides a cloud platform to offer on-demand services to complete consumers’ tasks, but assigning tasks to enterprises with different services requires many-to-many scheduling. The dynamic cloud environment puts forward higher requirements on scheduling algorithms’ real-time response and generalizability. Additionally, complex manufacturing tasks with flexible processing sequences also increase the decision-making difficulty. The existing approaches either have difficulty meeting the requirements of dynamics and fast-respond or struggle to effectively capture features of tasks with flexible processing sequences. To address these limitations, we develop a novel scheduling algorithm to solve a dynamic scheduling problem in the group service cloud manufacturing environment. Our proposal is formulated and trained by multi-agent reinforcement learning. The graph convolution network encodes tasks’ graph-structure features, and the recurrent neural network records each task’s processing trajectories. We independently design the action space and the reward function and train the algorithm with a mixing network under the centralized training decentralized execution architecture. Multi-agent reinforcement learning and graph convolution networks are rarely used to cloud manufacturing scheduling problems. Contrast experiments on a case study indicate that our proposal outperforms the other six multi-agent reinforcement learning-based scheduling algorithms in terms of scheduling performance and generalizability.

Automatic Seizure Identification from EEG Signals Based on Brain Connectivity Learning.

Epilepsy is a neurological disorder caused by brain dysfunction, which could cause uncontrolled behavior, loss of consciousness and other hazards. Electroencephalography (EEG) is an indispensable auxiliary tool for clinical diagnosis. Great progress has been made by current seizure identification methods. However, the performance of the methods on different patients varies a lot. In order to deal with this problem, we propose an automatic seizure identification method based on brain connectivity learning. The connectivity of different brain regions is modeled by a graph. Different from the manually defined graph structure, our method can extract the optimal graph structure and EEG features in an end-to-end manner. Combined with the popular graph attention neural network (GAT), this method achieves high performance and stability on different patients from the CHB-MIT dataset. The average values of accuracy, sensitivity, specificity, F1-score and AUC of the proposed model are 98.90%, 98.33%, 98.48%, 97.72% and 98.54%, respectively. The standard deviations of the above five indicators are 0.0049, 0.0125, 0.0116 and 0.0094, respectively. Compared with the existing seizure identification methods, the stability of the proposed model is improved by 78-95%.

Bot-MGAT: A Transfer Learning Model Based on a Multi-View Graph Attention Network to Detect Social Bots

Twitter, as a popular social network, has been targeted by different bot attacks. Detecting social bots is a challenging task, due to their evolving capacity to avoid detection. Extensive research efforts have proposed different techniques and approaches to solving this problem. Due to the scarcity of recently updated labeled data, the performance of detection systems degrades when exposed to a new dataset. Therefore, semi-supervised learning (SSL) techniques can improve performance, using both labeled and unlabeled examples. In this paper, we propose a framework based on the multi-view graph attention mechanism using a transfer learning (TL) approach, to predict social bots. We called the framework ‘Bot-MGAT’, which stands for bot multi-view graph attention network. The framework used both labeled and unlabeled data. We used profile features to reduce the overheads of the feature engineering. We executed our experiments on a recent benchmark dataset that included representative samples of social bots with graph structural information and profile features only. We applied cross-validation to avoid uncertainty in the model’s performance. Bot-MGAT was evaluated using graph SSL techniques: single graph attention networks (GAT), graph convolutional networks (GCN), and relational graph convolutional networks (RGCN). We compared Bot-MGAT to related work in the field of bot detection. The results of Bot-MGAT with TL outperformed, with an accuracy score of 97.8%, an F1 score of 0.9842, and an MCC score of 0.9481.

Point-of-Interest Recommendation Model Based on Graph Convolutional Neural Network

With the development of location-based social networks, the point-of-interest recommendation has become one of the research hotspots in the field of recommendation. However, traditional technologies like collaborative filtering are limited by the influence of data sparsity and cannot accurately capture the users’ preferences from the complex context. In order to address this problem, a recommendation model based on graph convolutional neural network is proposed, named RMGCN. RMGCN is composed of three parts: graph structure features extraction module, geographical factor evaluation module, and score calculation module. The graph structure feature extraction module is used to extract node features from the graph structure data composed of user check-in records. The geographical factor evaluation module is used to calculate the influence coefficient of geographical factors on user’s decision-making behaviors. The score calculation module is used to combine the output results of the above two modules and calculate the user’s preference scores of point-of-interests combined with temporal context and spatial context. Experimental results on two real-world datasets show that RMGCN has better recommendation performance than baselines.