Complex information network analysis is a vital research area in information diffusion, business marketing, personalized recommendation, and social influence analysis, etc. The complex information networks we focus on in this Special Issue refer to dynamic, heterogeneous, attribute, online, and/or direct information networks. In heterogeneous information networks, nodes are linked with multiple different types of edges, and nodes represent different types of entities (e.g., bus stations and metro stations in public transportation networks). In dynamic information networks, the topology of those networks changes over time as nodes/edges are added or removed. In addition, edges are always directed, and nodes are characterized by multiple attributes. How to mine behavioural features and rules from complex information networks is essential to related applications. Pairwise relations can only provide insights about local neighbourhoods and might not infer global hierarchical network structures, which is crucial for complex networks. Therefore, we require refined methods and feature extraction algorithms to enhance the prediction accuracy of complex systems, due to the complexity and diversity of the network data. How to design effective network representations that are capable of preserving hierarchical structures of networks is a promising direction for further work.
Learning a deep structure representation for complex information networks aims to project a graph into a low-dimensional vector space. Data mining and machine learning methods can easily deal with the network representation for further applications, such as link prediction, node classification, anomaly detection, and community detection. Due to the complexity of information networks, designing a novel network representation to deal with the heterogeneity and evolution of networks is a challenging and promising topic.
The aim of this Special Issue is to solicit contributions to fundamental research in deep structure representation and learning for complex information networks. Submissions discussing and introducing new algorithmic foundations and represen- tation formalisms are also welcome. We also seek studies on network representation applied to business, sociology, biology, health, and other industrial applications of complex information networks that help to solve real-world problems. Review articles discussing the current state of the art are also welcomed.
Potential topics include but are not limited to the following:
- Graph embedding
- Graph neural networks
- Generative adversarial net
- Deep neural networks for social networks Online social network analysis
- Random walk-based algorithms
- Matrix factorization for social networks Social influence analysis
- Link prediction
- Node classification
- Social anomaly detection
- Community detection
- Social recommender systems
- Knowledge graph learning
- Graph generation
- Network dynamics
- Heterogeneous social network analysis Subgraph structure learning
Ref: Deep Learning for Community Detection: Progress, Challenges and Opportunities, Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence Survey track. Pages 4981-4987. https://doi.org/10.24963/ijcai.2020/693 [PDF: https://www.ijcai.org/Proceedings/2020/0693.pdf]
Authors can submit their manuscripts through the Manuscript Tracking System at https://mts.hindawi.com/submit/journals/complexity/dsrlc/. Papers are published upon acceptance, regardless of the Special Issue publication date.
Lead Guest Editor
Jia Wu, Macquarie University, Sydney, Australia
jia.wu@mq.edu.au
Guest Editors
Chuan Zhou, Chinese Academy of Sciences,Beijing,China zhouchuan@amss.ac.cn
Shenghua Liu, Chinese Academy of Sciences, Beijing, China liushenghua@ict.ac.cn
Haishuai Wang, Harvard University, Boston, USA haishuai_wang@hms.harvard.edu
Submission Deadline
Friday, 9 October 2020
Publication Date
February 2021