Yixiang Fang is an Associate Professor in the School of Data Science, The Chinese University of Hong Kong, Shenzhen. He received Ph.D. in computer science from the University of Hong Kong in 2017. After that, he worked as a Research Associate in the School of Computer Science and Engineering, University of New South Wales, with Prof. Xuemin Lin. His research interests include querying, mining, and analytics of big graph data and big spatial data. He has published extensively in the areas of database and data mining, and most of his papers were published in top-tier conferences (e.g., PVLDB, SIGMOD, ICDE, NeurIPS, and IJCAI) and journals (e.g., TODS, VLDBJ, and TKDE), including One of the Best Papers in SIGMOD 2020. He received the 2021 ACM SIGMOD Research Highlight Award. He is an editorial board member of the journal of Information \& Processing Management (IPM). He has also served as program committee members for several top conferences (e.g., ICDE, KDD, AAAI, and IJCAI) and invited reviewers for top journals (e.g., TKDE, VLDBJ, and TOC) in the areas of database and data mining.

This SpringerBrief provides the first systematic review of the existing works of cohesive subgraph search (CSS) over large heterogeneous information networks (HINs). It also covers the research breakthroughs of this area, including models, algorithms and comparison studies in recent years. This SpringerBrief offers a list of promising future research directions of performing CSS over large HINs.

The authors first classify the existing works of CSS over HINs according to the classic cohesiveness metrics such as core, truss, clique, connectivity, density, etc., and then extensively review the specific models and their corresponding search solutions in each group. Note that since the bipartite network is a special case of HINs, all the models developed for general HINs can be directly applied to bipartite networks, but the models customized for bipartite networks may not be easily extended for other general HINs due to their restricted settings. The authors also analyze and compare these cohesive subgraph models (CSMs) and solutions systematically. Specifically, the authors compare different groups of CSMs and analyze both their similarities and differences, from multiple perspectives such as cohesiveness constraints, shared properties, and computational efficiency. Then, for the CSMs in each group, the authors further analyze and compare their model properties and high-level algorithm ideas.

This SpringerBrief targets researchers, professors, engineers and graduate students, who are working in the areas of graph data management and graph mining. Undergraduate students who are majoring in computer science, databases, data and knowledge engineering, and data science will also want to read this SpringerBrief.

1. Introduction