Temporal Network Embedding via Tensor Factorization

TL;DR

This paper introduces Toffee, a tensor decomposition-based method for learning embeddings of temporal networks that effectively capture evolving patterns and outperform existing approaches in link prediction tasks.

Contribution

The paper presents Toffee, a novel tensor factorization approach that encodes cross-time information to better model temporal network dynamics.

Findings

Toffee outperforms existing methods in link prediction on real-world temporal networks.

The tensor-tensor product operator effectively captures periodic changes in evolving networks.

Experimental results validate the superiority of Toffee in temporal network embedding tasks.

Abstract

Representation learning on static graph-structured data has shown a significant impact on many real-world applications. However, less attention has been paid to the evolving nature of temporal networks, in which the edges are often changing over time. The embeddings of such temporal networks should encode both graph-structured information and the temporally evolving pattern. Existing approaches in learning temporally evolving network representations fail to capture the temporal interdependence. In this paper, we propose Toffee, a novel approach for temporal network representation learning based on tensor decomposition. Our method exploits the tensor-tensor product operator to encode the cross-time information, so that the periodic changes in the evolving networks can be captured. Experimental results demonstrate that Toffee outperforms existing methods on multiple real-world temporal networks in generating effective embeddings for the link prediction tasks.