The Alpha Lab @ Dalian University of Technology
Explore the Unknown World Let's make a difference in people's lives.

Scientific Collaboration Dynamics

Sterling – Premium Website Template

Collaboration between researchers is recognized as being important in modern academia. The reason of is that it can offer lots of benefits for both academic research as well as scientific researches. In the sense that, how to exploit valuable information from collaboration network is becoming more and more attractive and meaningful.

Our research mainly focuses on predicting and discovering the patterns and influencing factors of academic success, generally the science of success in science. The most anticipated goal of our research includes: (1) Discovering the success spectrum of scholarly entities, in other words, building the academic DNA, (2) Constructing a comprehensive and scientific evaluation system, and (3) Predicting the future success.

So exciting, in the above mentioned filed, we have finished several achievements, such as (1) conference closure based on the triadic closure and focal closure (JCDL’16, Can Academic Conferences Promote Research Collaboration?). (2) deep learning model to infer advisor-advisee relationships based on dynamic scientific collaboration networks (JCDL’16, Mining Advisor-Advisee Relationships in Scholarly Big Data: A Deep Learning Approach).(3) exploring whether the sustainability of scientific collaboration including collaboration duration and collaboration times can be predicted(IW3C2’17, Is Scientific Collaboration Sustainability Predictable?).





Link Group

Sterling – Premium Website Template

The large, complex networks can be considered as structures whose nodes represent entities embedded in different context, and edges represent interactions between entities. Exploring the inner structure and identifying the mechanisms by which interactions evolve is a fundamental question that is still not well-understood. Higher order organizations in networks, i.e., motifs, have recently drawn a lot of attention.

Our research mainly focuses on mining, analyzing, and predicting interactions in multiple networks (e.g., academic networks, social network, and protein networks). We try to seek answers for following challenging questions from Big Network perspective: (1) How to mining the implicit interactions with limited information in the dynamic networks? (2) How to efficiently analyze the networks in higher order structures from novel viewpoints? (3) How to understand, profile, and analyze motifs in networks?

We develop approaches to infer advisor-advisee relationships based on dynamic scientific collaboration networks(WWW’17, Shifu: Deep Learning Based Advisor-advisee Relationship Mining in Scholarly Big Data). We are making efforts in handling especially high computational complexity brought by large-scale networks. Moreover, we aiming at proposing universal methods to solve network-based issues.





Introduction of Knowledge Group

Sterling – Premium Website Template

Knowledge is a familiarity, awareness, or understanding of someone or something, such as facts, information, descriptions, or skills. The evolutionary history of knowledge reflects the process of human to understand the world. Recently, data science revolutionizes everything, providing us the unprecedented ability to uncover the law between knowledge and human behavior.

Against this general background, we mainly focus on knowledge mining from two angles: education and knowledge graph. On one hand, education, as the specialized activity to deliver knowledge, contains tremendous knowledge-related information. We collect education data from various sources and explore student behavior from three meaningful aspects: academic performance, psychological condition and career development. On the other hand, as a crux graph-based description of the entities in the real world, knowledge graph raises a surge attention in the fields of network science and data science. Three directions we devoted in knowledge graph include the construction, representation and inference.





Mobile Computing in the Era of Big Data

Sterling – Premium Website Template

Mobile computing in the era of big data, generally enabled by resource-limited mobile devices, focuses on fostering the rise of updated self-configuring integrated computing-communication platforms to facilitate the real-time offloading and processing of big data streams, which is. It can be viewed as the convergence of Internet-based mobile networking and high-efficiency parallel computing.

Our research mainly focuses on the fundamental structure of mobile networks, human behavioural analysis and high-efficiency computing. We first propose a community-partition aware replica allocation method by exploiting social relationships(TPDS’14, Exploiting Social Relationship to Enable Efficient Replica Allocation in Ad-hoc Social Networks). Then, we propose a community-based load balancing method to improve the network performance by considering interest similarity and filter replication(TC’15, Community-Based Event Dissemination with Optimal Load Balancing). After that, we develop a distributed data information diffusion mechanism in mobile networks (TMC’16, PIS: A Multi-dimensional Routing Protocol for Socially-aware Networking). Furthermore, we investigate human behaviors and propose a copy adjustable incentive scheme to stimulate selfish nodes to cooperate in data forwarding (TVT’16, CAIS: A Copy Adjustable Incentive Scheme in Community-based Socially-Aware Networking). After that, the content dissemination architecture of vehicular social Networks is investigated (COMMAG’17, Vehicular Social Networks: Enabling Smart Mobility).





Science of Success in Science

Sterling – Premium Website Template

In academia, researchers tend to conduct their research following the directions of successful scholars or influential articles. But why similar people often experience diverging trajectories of accomplishment? Some scholars have experienced the accumulated long strings of success while others have encountered repeated failure. In other words, what factors contribute to determine the success still remain to be explored.

Our research mainly focuses on predicting and discovering the patterns and influencing factors of academic success, generally the science of success in science. The most anticipated goal of our research includes: (1) Discovering the success spectrum of scholarly entities, in other words, building the academic DNA, (2) Constructing a comprehensive and scientific evaluation system, and (3) Predicting the future success.

We proposed a novel algorithm to find academic rising stars according to their early academic records(WWW'15, CocaRank: A Collaboration Caliber-based Method for Finding Academic Rising Stars). We also evaluated the scientific impact of articles based on the conflict of interest relationships among scholars (PLOS ONE'16, Identifying Anomalous Citations for Objective Evaluation of Scholarly Article Impact).





Science of Scientific Team Science

Sterling – Premium Website Template

Scientific collaboration and teamwork continue to influence both practice of science and the production of knowledge. Nowadays, they have attracted increasing attention in various disciplines. As the network structure is becoming more and more complex, we have shifted our basic research unit from individual scholars to team-based scholars.

Our research mainly focuses on the problems of the Science of Scientific Team Science (SSTS). Our strong interest can be seen within the following issues from Big Network perspective: (1) How to define a new basic research unit in big scientific network? (2) How to optimize team structure? (3) How to become successful teams?




Scholarly Recommendation

Sterling – Premium Website Template

Scholarly data is getting bigger and bigger, which leads to the problem of academic information overload. Scholarly recommendation aims to suggest items of potential interests for solving academic information overload.

Our goal is to provide various personalized academic recommendation services including (1) collaborator recommendation, (2) article recommendation, and (3) venue recommendation.

We proposed a socially-aware recommendation system to recommend venues and sessions for conference participants (UIC’13, Socially-Aware Venue Recommendation for Conference Participants, Best Paper Award). We also proposed to recommend most valuable collaborators with a biased random walk considering link importance (TETC’14, MVCWalker: Random Walk Based Most Valuable Collaborators Recommendation Exploiting Academic Factors) and further exploited publication contents and collaboration networks for potential collaborator recommendation (PLOS ONE’16, Exploiting Publication Contents and Collaboration Networks for Collaborator Recommendation). We proposed to recommend scientific articles based on common author relations and historical preferences (TBD’16, Scientific Article Recommendation: Exploiting Common Author Relations and Historical Preferences).





Big Trajectory Data

CS2 Group

 In recent years, with the pervasive application of intelligent transportation equipments such as GPS devices, traffic cameras, smart cards, and traffic sensors, big trajectory data are more easily collected than before. Analyzing and mining the patterns hidden in big trajectory data has been a hot research field associated with transportation management, urban planning, epidemic control, mobile platform application, and so on.

Our research mainly focuses on converging multi-source heterogeneous big trajectory data, discovering implicit human mobility patterns, mining social functions in different city areas, and generating private car dataset. In addition, we are very interested in mining the potential relationship between human social behavior and economic sociology in big traffic scenario.

We propose a novel approach to estimate and predict the urban traffic congestion utilizing fuzzy evaluation and particle swarm optimization algorithm (UIC'15, Taxi Operation Optimization Based on Big Traffic Data). We also propose a data-driven taxi operation strategy to maximize drivers’ profit, reduce energy consumption, and decrease environment pollution. Specifically, we introduce the Time-Location-Sociality model which can identify three dimensional properties of city dynamics to predict the number of passengers in different social functional regions (UIC'15, Taxi Operation Optimization Based on Big Traffic Data).





Socially-Aware Networking

Sterling – Premium Website Template

Socially-Aware Networking (SAN) is a new communication paradigm, in which the social characteristics of mobile nodes are exploited to improve the performance of data distribution. The motivation is that the nodes’ social ties have long-term characteristics that can help build stable connectivity graphs among them.

Our research mainly focuses on developing distributed data routing protocols (TMC’16, PIS: A Multi-dimensional Routing Protocol for Socially-aware Networking) and dissemination architectures (COMMAG’17, Vehicular Social Networks: Enabling Smart Mobility) in SANs.

We proposed a community-partition aware replica allocation method by exploiting social relationship (TPDS’14, Exploiting Social Relationship to Enable Efficient Replica Allocation in Ad-hoc Social Networks). Then, we proposed a community-based load balancing method to improve the network performance by clustering brokers in a community by considering interest similarity and filter replication (TC’15, Community-Based Event Dissemination with Optimal Load Balancing). Furthermore, we proposed a copy adjustable incentive scheme to stimulate selfish nodes to cooperate in data forwarding (TVT’16, CAIS: A Copy Adjustable Incentive Scheme in Community-based Socially-Aware Networking).





Research Areas


  • * Big Data
  • * Computational Social Science
  • * Social Computing
  • * Mobile Computing