Network Structure Information Research Articles

Global stabilizing control of large-scale biomolecular regulatory networks.

Cellular behavior is determined by complex non-linear interactions between numerous intracellular molecules that are often represented by Boolean network models. To achieve a desired cellular behavior with minimal intervention, we need to identify optimal control targets that can drive heterogeneous cellular states to the desired phenotypic cellular state with minimal node intervention. Previous attempts to realize such global stabilization were based solely on either network structure information or simple linear dynamics. Other attempts based on non-linear dynamics are not scalable. Here, we investigate the underlying relationship between structurally identified control targets and optimal global stabilizing control targets based on non-linear dynamics. We discovered that optimal global stabilizing control targets can be identified by analyzing the dynamics between structurally identified control targets. Utilizing these findings, we developed a scalable global stabilizing control framework using both structural and dynamic information. Our framework narrows down the search space based on strongly connected components and feedback vertex sets then identifies global stabilizing control targets based on the canalization of Boolean network dynamics. We find that the proposed global stabilizing control is superior with respect to the number of control target nodes, scalability, and computational complexity. We provide a GitHub repository that contains the DCGS framework written in Python as well as biological random Boolean network datasets (https://github.com/sugyun/DCGS). Supplementary data are available at Bioinformatics online.

Community Discovery Algorithm Based on Multi-Relationship Embedding

Complex systems in the real world often can be modeled as network structures, and community discovery algorithms for complex networks enable researchers to understand the internal structure and implicit information of networks. Existing community discovery algorithms are usually designed for single-layer networks or single-interaction relationships and do not consider the attribute information of nodes. However, many real-world networks consist of multiple types of nodes and edges, and there may be rich semantic information on nodes and edges. The methods for single-layer networks cannot effectively tackle multi-layer information, multi-relationship information, and attribute information. This paper proposes a community discovery algorithm based on multi-relationship embedding. The proposed algorithm first models the nodes in the network to obtain the embedding matrix for each node relationship type and generates the node embedding matrix for each specific relationship type in the network by node encoder. The node embedding matrix is provided as input for aggregating the node embedding matrix of each specific relationship type using a Graph Convolutional Network (GCN) to obtain the final node embedding matrix. This strategy allows capturing of rich structural and attributes information in multi-relational networks. Experiments were conducted on different datasets with baselines, and the results show that the proposed algorithm obtains significant performance improvement in community discovery, node clustering, and similarity search tasks, and compared to the baseline with the best performance, the proposed algorithm achieves an average improvement of 3.1% on Macro-F1 and 4.7% on Micro-F1, which proves the effectiveness of the proposed algorithm.

A Two-Stage Deep-Learning Model for Link Prediction Based on Network Structure and Node Attributes

Link prediction, which is used to identify the potential relationship between nodes, is an important issue in network science. In existing studies, the traditional methods based on the structural similarity of nodes make it challenging to complete the task of link prediction in large-scale or sparse networks. Although emerging methods based on deep learning can solve this problem, most of the work mainly completes the link prediction through the similarity of the representation vector of network structure information. Many empirical studies show that link formation is affected by node attributes, and similarity is not the only criterion for the formation of links in reality. Accordingly, this paper proposed a two-stage deep-learning model for link prediction (i.e, TDLP), where the node representation vector of the network structure and attributes was obtained in the first stage, while link prediction was realized through supervised learning in the second stage. The empirical results on real networks showed that our model significantly outperforms the traditional methods (e.g., CN and RA), as well as newly proposed deep-learning methods (e.g., GCN and VGAE). This study not only proposed a deep-learning framework for link prediction from the perspective of structure and attribute fusion and link distribution capture, but also lays a methodological foundation for practical applications based on link prediction.

Identification of Important Nodes in Multilayer Heterogeneous Networks Incorporating Multirelational Information

Centrality is an effective method to identify important nodes in complex networks, but it is still a challenge to find influential nodes by making full use of multiple relationships and global network topological features in complex networks. To address these problems, this article proposes an importance identification method for multilayer heterogeneous network node by incorporating multirelational information (MLC). This method studies the relational characteristics of heterogeneous nodes in detail and divides the heterogeneous nodes into different layers according to the node types, which can be further divided into core and auxiliary layers. The importance of the auxiliary layer is quantified by designing the interlayer influence and determining the interlayer influence weights of different connectivity influences; the centrality score of heterogeneous nodes under multiconnectivity relationships is fused using the transmission characteristics of internode relationships in the auxiliary layer, which in turn measures the importance of nodes in the core layer. To evaluate the proposed algorithm, we conduct experiments on five real multilayer heterogeneous networks of different sizes. The results show that MLC can make full use of different types of internode association relationship information, effectively fuse network structure information such as the neighbor weights of core and auxiliary layer nodes, and outperform the existing techniques in identifying important nodes.

Influence maximization through exploring structural information

Influence maximization (IM) is a widely investigated issue in the study of social networks because of its potential commercial and social value. The purpose of IM is to identify a group of influential nodes that will spread information to other nodes in a network while simultaneously maximizing the number of nodes that are ultimately influenced. Traditional IM methods have different limitations, such as limited scalability to address large-scale networks and the neglect of community structural information. Here, we propose a novel influence maximization approach, i.e., the layered gravity bridge algorithm (LGB), to address the IM problem, which emphasizes the local structural information of networks and combines community detection algorithms with an improved gravity model. With the proposed LGB, a community detection method is used to derive the communities, and then the bridge nodes are found, which can be regarded as possible candidate seeds. Later, communities are merged into larger communities according to our proposed algorithm, and new bridge nodes are determined. Finally, all candidate seed nodes are sorted through an improved gravity model to determine the final seed nodes. The algorithm fully explores the network structural information provided by the communities, thereby making it superior to the current algorithms in terms of the number of ultimately infected nodes. Furthermore, our proposed algorithm possesses the potential to alleviate the influence overlap effect of seed nodes. To verify the effect of our approach, the classical SIR model is adopted to propagate information with the selected seed nodes, while experiments are performed on several practical datasets. As indicated by the obtained results, the performance of our proposed algorithm outperforms existing ones.

Information network, public disclosure and asset prices

We propose in this paper a noisy rational expected equilibrium (NREE) model by taking both public and private information into account with an embedded information network structure among market traders. We derive closed-form expressions for five variables about market reaction and market quality as a function of the topological structure of the network, and we obtain several novel results. First, the information network directly affects the price discovery of private information, but it indirectly influences the price discovery of public information. Second, network connectedness negatively influences both price change and trading volume, while network uniformity only affects trading volume positively and does not impact price change. Third, network connectedness suppressed the marginal effect of public disclosure on market quality, i.e., the information sharing among traders weakens the market quality improvement caused by public disclosure. We extend the NREE model and document a crowding-out effect of information networks in the price discovery of public disclosure. The mechanism of the crowding-out effect also provides a theoretical interpretation of the under-reaction of public disclosure in the market.

Association prediction of CircRNAs and diseases using multi-homogeneous graphs and variational graph auto-encoder

As a non-coding RNA molecule with closed-loop structure, circular RNA (circRNA) is tissue-specific and cell-specific in expression pattern. It regulates disease development by modulating the expression of disease-related genes. Therefore, exploring the circRNA-disease relationship can reveal the molecular mechanism of disease pathogenesis. Biological experiments for detecting circRNA-disease associations are time-consuming and laborious. Constrained by the sparsity of known circRNA-disease associations, existing algorithms cannot obtain relatively complete structural information to represent features accurately. To this end, this paper proposes a new predictor, VGAERF, combining Variational Graph Auto-Encoder (VGAE) and Random Forest (RF). Firstly, circRNA homogeneous graph structure and disease homogeneous graph structure are constructed by Gaussian interaction profile (GIP) kernel similarity, semantic similarity, and known circRNA-disease associations. VGAEs with the same structure are employed to extract the higher-order features by the encoding and decoding of input graph structures. To further increase the completeness of the network structure information, the deep features acquired from the two VGAEs are summed, and then train the RF with sparse data processing capability to perform the prediction task. On the independent test set, the Area Under ROC Curve (AUC), accuracy, and Area Under PR Curve (AUPR) of the proposed method reach up to 0.9803, 0.9345, and 0.9894, respectively. On the same dataset, the AUC, accuracy, and AUPR of VGAERF are 2.09%, 5.93%, and 1.86% higher than the best-performing method (AEDNN). It is anticipated that VGAERF will provide significant information to decipher the molecular mechanisms of circRNA-disease associations, and promote the diagnosis of circRNA-related diseases.

Dynamic Community Detection Based on Evolutionary DeepWalk

To fully characterize the evolution process of the topological structure of dynamic communities, we propose a dynamic community detection based on Evolutionary DeepWalk (DEDW) for the high-dimensional data and dynamic characteristics. First, DEDW solves the problem of data sparseness in the process of dynamic network data representation through graph embedding. Then, DEDW uses the DeepWalk algorithm to generate node embedding feature vectors based on the characteristics of the stable change of the community structure; finally, DEDW integrates historical network structure information to generate evolutionary graph features and implements dynamic community detection with the K-means algorithm. Experiments show that DEDW can mine the time-smooth change characteristics of dynamic communities, solve the problem of data sparseness in the process of node embedding, fully consider historical structure information, and improve the accuracy and stability of dynamic community detection.

Learning graph deep autoencoder for anomaly detection in multi-attributed networks

Anomaly detection in multi-attributed networks has become increasingly important and has significant implications in various domains, such as intrusion detection, botnet detection, financial fraud detection and event detection. However, detecting rare anomalous nodes is always challenging in a multi-attributed network with a large set of data observations without labels. In this work, we address this problem by learning a graph deep autoencoder framework named as GDAE. The GDAE first jointly takes both the network structure and node attributes as input to calculate the embedding for every node in multi-attributed networks using a deep attention mechanism by attending to its neighbor nodes. Meanwhile, through multiple layers of nonlinear transformations, the GDAE efficiently captures the non-linearity of data and the complex interactions of both node attribute information and network structure information. GDAE then focuses on conducting the decoders of both network structure and node attributes based on the learned node embedding. Subsequently, the reconstruction errors based on the aforementioned encoder and decoder operations are employed to discover the anomalies in multi-attributed networks. Extensive experiments using four real-world datasets from different domains demonstrate that our approach performs superior over representative baseline approaches.

GA-Net: A geometry prior assisted neural network for road extraction

With geospatial intelligence research developing rapidly, automatic road extraction is becoming a fundamental and challenging task. Due to the special geometric structure and spectral information of road networks, existing methods suffer from incomplete and fractured results. In this work, a novel road extraction convolutional neural network, incorporating the road boundary details and road junction information via a dual-branch multi-task structure, is proposed to learn synergistic feature representations and strengthen road connectivity. Firstly, a BiFPN-based feature aggregation module is utilised to bridge the semantic gap between low-level and high-level feature maps, allowing multi-scale spatial details to be fully fused. Secondly, the boundary auxiliary branch, using a U-shaped network with a spatial-channel attention module, captures residential information for the backbone to enhance the subtleties of road edges. Thirdly, the node inferring branch models the road junction position jointly with the road surface, aiming to strengthen the topology structure and reduce the fragmented road segments. We perform experiments on three diverse road datasets, namely the DeepGlobe dataset, Massachusetts dataset, and SpaceNet dataset. The results demonstrate that our model shows an overall performance improvement over some SOTA algorithms and the IoU indicator achieves 3.86%, 0.79%, and 1.71% improvements over Unet on the three datasets, respectively. • A novel Geometric Prior Assisted Neural Network (GA-Net) is proposed for automatic road extraction. The boundary information is introduced to the model, which strengthens the geometric details of road edges. The road junction information is jointly learned with the road surface information, which reduces the missing gap between road segments. • A feature fusion module is designed to bridge the semantic gap between multi-scale features maps. An attention module is designed to strengthen the road geometric information. The two light modules can be easily ensemble with other models. • We achieves better results than some SOTA algorithms for three popular datasets. And results of our model shows better completeness and connectivity.

Cross-domain entity identity association analysis and prediction based on representation learning

Cross-domain identity association of network entities is a significant research challenge and a vital issue of practical value in relationship discovery and service recommendation between things in the Internet of things, cyberspace resources surveying mapping, threat tracking, and intelligent recommendation. This task usually adds additional difficulty to the research in practical applications due to the need to link across multiple platforms. The existing entity identity association methods in cross-domain networks mainly use the attribute information, generated content, and network structure information of network user entities but do not fully use the inherent strong positioning characteristics of active nodes in the network. In this article, we analyzed the structural characteristics of existing relational networks. We found that the hub node has the role of identity association positioning, and the importance of identity association reflected by different nodes is different. Moreover, we creatively designed a network representation learning method. We proposed a supervised learning identity association model combined with a representation learning method. Experiments on the public data set show that using the identity association method proposed in this article, the ranking accuracy of user entity association similarity is about 30% and 25% higher than the existing two typical methods.

IPiDA-GCN: Identification of piRNA-disease associations based on Graph Convolutional Network.

Piwi-interacting RNAs (piRNAs) play a critical role in the progression of various diseases. Accurately identifying the associations between piRNAs and diseases is important for diagnosing and prognosticating diseases. Although some computational methods have been proposed to detect piRNA-disease associations, it is challenging for these methods to effectively capture nonlinear and complex relationships between piRNAs and diseases because of the limited training data and insufficient association representation. With the growth of piRNA-disease association data, it is possible to design a more complex machine learning method to solve this problem. In this study, we propose a computational method called iPiDA-GCN for piRNA-disease association identification based on graph convolutional networks (GCNs). The iPiDA-GCN predictor constructs the graphs based on piRNA sequence information, disease semantic information and known piRNA-disease associations. Two GCNs (Asso-GCN and Sim-GCN) are used to extract the features of both piRNAs and diseases by capturing the association patterns from piRNA-disease interaction network and two similarity networks. GCNs can capture complex network structure information from these networks, and learn discriminative features. Finally, the full connection networks and inner production are utilized as the output module to predict piRNA-disease association scores. Experimental results demonstrate that iPiDA-GCN achieves better performance than the other state-of-the-art methods, benefitted from the discriminative features extracted by Asso-GCN and Sim-GCN. The iPiDA-GCN predictor is able to detect new piRNA-disease associations to reveal the potential pathogenesis at the RNA level. The data and source code are available at http://bliulab.net/iPiDA-GCN/.

Structural centrality of networks can improve the diffusion-based recommendation algorithm

The recommendation system has become an indispensable information technology in the real world. The recommendation system based on the diffusion model has been widely used because of its simplicity, scalability, interpretability, and many other advantages. However, the traditional diffusion-based recommendation model only uses the nearest neighbor information, which limits its efficiency and performance. Therefore, in this article, we introduce the centralities of complex networks into the diffusion-based recommendation system and test its performance. The results show that the overall performance of heat conduction algorithm can be improved by 184%–280%, using the centrality of complex networks, reaching almost the same accuracy level as the mass diffusion algorithm. Therefore, the recommendation system combining the high-order network structure information is a potentially promising research direction in the future.

Unidirectional and bidirectional LSTM models for edge weight predictions in dynamic cross-market equity networks

Predicting financial networks has important implications in finance. However, less research attention has been given in this direction. This study aims to predict cross market linkage strengths in financial networks using dynamic edge weight prediction. The study builds edge weight prediction models using deep learning based unidirectional and bidirectional long short-term memory (LSTM and BiLSTM) approaches. The models are built on temporally variant equity cross market networks in Asia. A rolling window approach is used to generate temporally synchronous observations of edge weight structures and subsequent networks. The models are trained with a series of historical network structure information. Tuning of hyperparameters was performed to obtain the optimized models. The models were validated using nested cross validation methods. The applicability of the optimized models was assessed for different scenarios, and promising results were found for edge weight predictions on both unfiltered and filtered structures. The study also shows that the predictive performance of the BiLSTM model is the same across correlated (positively weighted) and anti-correlated (negatively weighted) edges. However, for LSTM the same does not hold true. The results also demonstrate that the proposed models predict edge weights better during normal conditions than in the crisis period. The proposed models are benchmarked against ARIMA and RNN. To our knowledge, this is the first attempt to build network prediction models for cross market equity networks. The findings can have key implications such as managing international portfolio diversifications and controlling systemic risk transmissions.

ST-AGRNN: A Spatio-Temporal Attention-Gated Recurrent Neural Network for Traffic State Forecasting

Accurate traffic state prediction plays an important role in traffic guidance, travel planning, etc. Due to the existence of complex spatio-temporal relationships, there are some challenges in forecasting. Firstly, in terms of spatial correlation, some models only consider the road network structure information, and ignore the relative location relationships between nodes. Secondly, some models ignore the different impacts of nodes in the global road network on traffic. To solve these problems, we propose a new traffic state-forecasting model, namely, spatio-temporal attention-gated recurrent neural network (ST-AGRNN). In the proposed model, structure-based and location-based localized spatial features are obtained simultaneously by Graph Convolutional Networks (GCNs) and DeepWalk. The localized temporal features are obtained by gated recurrent unit (GRU). The attention-based approach is used to obtain global spatio-temporal features. Experimental validation is performed with two real-world public datasets, and the results show that the ST-AGRNN model outperforms the state-of-the-art methods.

Detecting COVID-19 vaccine hesitancy in India: a multimodal transformer based approach.

COVID-19 has emerged as the greatest threat in recent times, causing extensive mortality and morbidity in the entire world. India is among the highly affected countries suffering severe disruptions due this pandemic. To overcome the adverse effects of COVID-19, vaccination has been identified as the most effective preventive measure globally. However, a growing amount of hesitancy has been observed among the general public regarding the efficacy and possible side-effects of vaccination. Such hesitancy may proved to be the greatest hindrance towards combating this deadly pandemic. This paper introduces a multimodal deep learning method for Indian Twitter user classification, leveraging both content-based and network-based features. To explore the fundamental features of different modalities, improvisations of transformer models, BERT and GraphBERT are utilized to encode the textual and network structure information. The proposed approach thus integrates multiple data representations, utilizing the advances in both transformer based deep learning as well as multimodal learning. Experimental results demonstrates the efficacy of proposed approach over state of the art approaches. Aggregated feature representations from multiple modalities embed additional information that improves the classification results. The findings of the proposed model has been further utilized to perform a study on the dynamics of COVID-19 vaccine hesitancy in India.

INSISTC: Incorporating network structure information for single-cell type classification

Uncovering gene regulatory mechanisms in individual cells can provide insight into cell heterogeneity and function. Recent accumulated Single-Cell RNA-Seq data have made it possible to analyze gene regulation at single-cell resolution. Understanding cell-type-specific gene regulation can assist in more accurate cell type and state identification. Computational approaches utilizing such relationships are under development. Methods pioneering in integrating gene regulatory mechanism discovery with cell-type classification encounter challenges such as determine gene regulatory relationships and incorporate gene regulatory network structure. To fill this gap, we developed INSISTC, a computational method to incorporate gene regulatory network structure information for single-cell type classification. INSISTC is capable of identifying cell-type-specific gene regulatory mechanisms while performing single-cell type classification. INSISTC demonstrated its accuracy in cell type classification and its potential for providing insight into molecular mechanisms specific to individual cells. In comparison with the alternative methods, INSISTC demonstrated its complementary performance for gene regulation interpretation.

Weighted deep factorizing heterogeneous molecular network for genome-phenome association prediction.

Genome-phenome association (GPA) prediction can promote the understanding of biological mechanisms about complex pathology of phenotypes (i.e., traits and diseases). Traditional heterogeneous network-based GPA approaches overwhelmingly need to project heterogeneous data toward homogeneous network for data fusion and prediction, such projections result in the loss of heterogeneous network structure information. Matrix factorization based data fusion can avoid such projection by integrating multi-type data in a coherent way, but they typically perform linear factorization and cannot mine the nonlinear relationships between molecules, which compromise the accuracy of GPA analysis. Furthermore, most of them can not selectively synergy network topology and node attribution information in a principle way. In this paper, we propose a weighted deep matrix factorization based solution (WDGPA) to predict GPAs by selectively and differentially fusing heterogeneous molecular network and diverse attributes of nodes. WDGPA firstly assigns weights to inter/intra-relational data matrices and attribute data matrices, and performs deep matrix factorization on these matrices of heterogeneous network in a cooperative manner to obtain the nonlinear representations of different nodes. In addition, it performs low-rank representation learning on the attribute data with the shared nonlinear representations. In this way, both the network topology and node attributes are jointly mined to explore the representations of molecules and complex interplays between molecules and phenotypes. WDGPA then uses the representational vectors of gene and phenotype nodes to predict GPAs. Experimental results on maize and human datasets confirm that WDGPA outperforms competitive methods by a large margin under different evaluation protocols.

Implementation of Network Data Mining Algorithm for Associated Users Based on Multi-Information Fusion

In order to accurately and effectively mine relevant users in social networks, we can stop false information and illegal activities in the network, thereby ensuring the safety and integrity of the network environment. A method is proposed for the implementation of a data mining algorithm of a user network based on the fusion of several data. AUMA-MRL (associated user mining algorithm based on multi-information representation learning) proposes an associated user mining algorithm based on node characteristics, neighborhood information, and global network structure information. The steps of the algorithm are as follows: combining each user of the social network into a node using a method where each node is installed separately and combining network user characteristics and user relationship information. A user pair is a network similarity vector that represents the similarity of users in different dimensions. Based on these similarity vectors, a corresponding user separation algorithm is formed. It examines the feasibility and efficiency of the AUMA-MRL algorithm for researching relevant users. The proportion of associated users in the network to be fused is lower than that of nonassociated users, and the prediction has little effect on improving the recall rate of nonassociated users, so the recall rate is slightly lower than the accuracy. This algorithm can quickly get the embedding of new nodes and the similarity vector between new nodes and other nodes in the network, so as to quickly mine the associated users of new nodes in the network and enhance the robustness of the network associated user mining algorithm.

Fusing visual and textual content for knowledge graph embedding via dual-track model

Large-scale knowledge graphs are usually incomplete. Knowledge graph embedding has achieved encouraging performance in alleviating the incompleteness of knowledge graphs. There are approaches to leverage the multi-modal content, such as text description and images, to improve the performance of knowledge graph embedding. However,due to the heterogeneity across different modalities, current methods are not effective to fuse the multi-modal content and network structure information to learn the embedding. In this work, a dual-track model DuMF for knowledge graph embedding enhancement is proposed. The model includes two tracks to fuse multi-modal content and network structure respectively. In each track, the expressiveness of joint features is improved by the bilinear method, and meanwhile the task-specific important features are learned by deliberate attention. Finally, the fused features are generated by the gating network. To extensively evaluate the model, two challenging datasets are enriched with additional multi-modal data. Experimental results show that DuMF is superior to baselines in link prediction. The flexibility of the model is promising for the further improvement of model performance.