IFM Lab

Our lab focuses on fusing multiple large-scale information sources of diverse varieties together, and carrying out synergistic data mining tasks across these fused sources in one unified analytic.
Fusing and mining multiple information sources of large volumes and diverse varieties is a fundamental problem in big data studies.
We strive to develop general methodologies for information fusion and mining, which will be shown to work well for a diverse set of applications.

(Our lab has several positions avaiable. More information here.)

IFM.py Toolkit

IFM.py is an open-source toolkit released by IFM Lab. The toolkit covers the state-of-art algorithms for information fusion and data mining.

Fusion Learning Textbook

Fusion Learning is a textbook about informaiton fusion and data mining, mainly about the application on heterogeneous social networks.

Multiple Source Datasets

IFM Lab releases several multi-source datasets on Online Social Networks, Knowledge Graphs, Geographical Locations, Bio-Informatics, etc.

Online Social Media Analysis

To enjoy different kinds social network services, people nowadays are usually involved in multiple online social networks simultaneously. Fusion and mining multiple aligned online social networks is one of the main stream research project of IFM Lab, which involves research problems like network alignment, cross-network link prediction, community detection, information diffusion, viral marketing, and network embedding etc.

Multiple Modality Text Mining

Multiple language systems co-exist simultaneously in the real world, like various natural languages and programming languages. When referring to certain concepts, different language will have their unique corpus sets. Identifying the correspondence relationships between different lauguage corpus sets is a new direction to be explored in IFM Lab.

Heterogeneous Knowledge Graph Studies

Knowledge and concepts in the real-world can be represented as a network structured diagram outlining their intrinsic relationships, namely the knowledge graphs. Many knowledge and concepts are shared by different knowledge graphs at the same time. Fusion of knowledge graphs from different sources is important for cross-source synergistic knowledge discovery.

Enterprise Fusion and Knowledge Discovery

In modern enterprise, to facilitate the communication among employees, a new type of online soical networks have been launched inside the firewall of the companies, namely the enterprise social networks (ESNs). Effective fusion of enterprise internal information sources, like ESNs, organizational chart, employee profile information and collaboration notes, for knowledge discovery is an important research projects in IFM Lab.

Bio-Medical Research

About the patients, different types of information will be collected before determining the desease they may have, including the brain fMRI image, regular physical diagnosis, and various drug tests. These complementary informaiton sources will provide a comprehensive information about the patient. Effectively fusing these information sources will lead to more precise diagnosis results for the patients.

Geographical Data Sources Fusion

From the government and traffic service providers, different kinds of data about cities can be obtained, like crime rate, population density, bus drop-off points, metro train routes. Effective fusion of such diverse information sources together provides a new research direction for studying traffic and urban planning problems from the data driven perspective.