Network Representation Learning

logo

Lab of Media and Network

Department of Computer Science & Technology

Tsinghua University

Community Preserving Network Embedding

X Wang, P Cui, J Wang, J Pei, W Zhu, S Yang. AAAI 2017

M-NMF learns the representations of nodes in a network which can preserve the microscopic structure (first- and second-order proximities of nodes) and mesoscopic community structure.

Motivation

Previous work focuses on preserving the microscopic structure when they learn the representations of nodes, while the community structure, which is one of the most prominent feature of network, is largely ignored. M-NMF is network embedding method based on the framework of nonnegative matrix factorization (NMF). Specifically, it discovers the community structure by maximizing the modularity, and learns the representations of nodes by NMF. Meanwhile, M-NMF builds their consensus relationships, so that the community structure can guide the representations of nodes. Also, we derive the updating rules of M-NMF with correctness and convergence guarantees.

pic

Datasets

Links to datasets used in the paper:

References

Community Preserving Network Embedding. Xiao Wang, Peng Cui, Jing Wang, Jian Pei, Wenwu Zhu, Shiqiang Yang. Proceedings of the 31st AAAI Conference on Artificial Intelligence (AAAI). 2017.