DeepNC: Deep Generative Network Completion

Tran, Cong; Shin, Won-Yong; Spitz, Andreas; Gertz, Michael

doi:10.48550/arxiv.1907.07381

Cited by 6 publications

(7 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Link prediction, as the traditional task in network inference, tries to infer the lost links in network structure according to the linking patterns of existing connections 17,18 . Although numerous algorithms have been developed to complete the unobservable links of a large network with high accuracy 19,20 , all of these approaches require the complete node information but it is always unavailable in practice 13,21 . Link prediction cannot solve the inference problem under the condition that the network contains unobservable nodes.…”

Section: Introductionmentioning

confidence: 99%

“…Xu et al proposed a GCN based model, which regards the process of completing a graph as a network growth process, and learns the rules of the growth to complement the full network 29 . Tran et al solved the problem by training a graph generative model to learn the connection patterns among a large set of similar graphs and use these patterns to infer the missing connections 21 . All of these network completion methods depend on a partially observable network structure because they try to discover the latent patterns of the observable connections and to infer the unknown structures.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Inferring Network Structure with Unobservable Nodes from Time Series Data

Zhang¹,

Zhang²,

Du³

et al. 2022

Preprint

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Inferring Network Structure with Unobservable Nodes from Time Series Data

Zhang¹,

Zhang²,

Du³

et al. 2022

Preprint

View full text Add to dashboard Cite

“…Graph generation entails modeling and generating real-world graphs, and it has applications in several domains, such as understanding interaction dynamics in social networks [47,128,129], link prediction [70,113], and anomaly detection [109]. Owing to its many applications, the development of generative models for graphs has a rich history, resulting in famous models such as random graphs, small-world models, stochastic block models, and Bayesian network models, which generate graphs based on apriori structural assumptions [98].…”

Section: Introductionmentioning

confidence: 99%

A Systematic Survey on Deep Generative Models for Graph Generation

Guo¹,

Zhao²

2020

Preprint

View full text Add to dashboard Cite

Graphs are important data representations for describing objects and their relationships, which appear in a wide diversity of real-world scenarios. As one of a critical problem in this area, graph generation considers learning the distributions of given graphs and generating more novel graphs. Owing to its wide range of applications, generative models for graphs have a rich history, which, however, are traditionally hand-crafted and only capable of modeling a few statistical properties of graphs. Recent advances in deep generative models for graph generation is an important step towards improving the fidelity of generated graphs and paves the way for new kinds of applications. This article provides an extensive overview of the literature in the field of deep generative models for graph generation. Firstly, the formal definition of deep generative models for the graph generation as well as preliminary knowledge is provided. Secondly, two taxonomies of deep generative models for unconditional, and conditional graph generation respectively are proposed; the existing works of each are compared and analyzed. After that, an overview of the evaluation metrics in this specific domain is provided. Finally, the applications that deep graph generation enables are summarized and five promising future research directions are highlighted.

show abstract

“…Link prediction, as the traditional task in network science, tries to infer the lost links in network structure according to the linking patterns of existing connections [12,13]. Although numerous algorithms have been developed to complete the missing links of a large network with high accuracy [14,15], all of these approaches require the complete nodes information but it is always unavailable in practice [8,16]. link prediction cannot solve the inference problem under the condition that the network contains unobservable nodes.…”

Section: Introductionmentioning

confidence: 99%

“…Da Xu et al regard the complete network as the growth of the partial network [25], then they train the GCN to learn the growing process with the partially observed network and generalized to the complete unknown network structure. Cong Tran, Won-Yong Shin, and Andreas Spitz et al solve the problem by training a graph generating model to learn the connection patterns among a large set of similar graphs for training and applied the trained generating model to complete the missing information [16]. All of these network completion methods depend on a partially observed network structure because they try to discover the latent patterns of the observed connections and to infer the unknown structures.…”

Section: Introductionmentioning

confidence: 99%

Inference for Network Structure and Dynamics from Time Series Data via Graph Neural Network

Zhang¹,

Zhang²,

Du³

et al. 2020

Preprint

View full text Add to dashboard Cite

Network structures in various backgrounds play important roles in social, technological, and biological systems. However, the observable network structures in real cases are often incomplete or unavailable due to the measurement errors or private protection issues. Therefore, inferring the complete network structure is useful for understanding complex systems. The existing studies have not fully solved the problem of inferring network structure with partial or no information about connections or nodes. In this paper, we tackle the problem by utilizing time series data generated by network dynamics. We regard the network inference problem based on dynamical time series data as a problem of minimizing errors for predicting future states and proposed a novel data-driven deep learning model called Gumbel Graph Network (GGN) to solve the two kinds of network inference problems: Network Reconstruction and Network Completion. For the network reconstruction problem, the GGN framework includes two modules: the dynamics learner and the network generator. For the network completion problem, GGN adds a new module called the States Learner to infer missing parts of the network. We carried out experiments on discrete and continuous time series data. The experiments show that our method can reconstruct up to 100% network structure on the network reconstruction task. While the model can also infer the unknown parts of the structure with up to 90% accuracy when some nodes are missing. And the accuracy decays with the increase of the fractions of missing nodes. Our framework may have wide application areas where the network structure is hard to obtained and the time series data is rich.

show abstract

DeepNC: Deep Generative Network Completion

Cited by 6 publications

References 33 publications

Inferring Network Structure with Unobservable Nodes from Time Series Data

Inferring Network Structure with Unobservable Nodes from Time Series Data

A Systematic Survey on Deep Generative Models for Graph Generation

Inference for Network Structure and Dynamics from Time Series Data via Graph Neural Network

Contact Info

Product

Resources

About