Gaussian Embedding of Temporal Networks – Modeling Network Interactions over Time with Explicit Uncertainty

Raphaël Romero from IDLab-AIDA created a model for interactions between entities that take place over time, a data structure also known as a temporal network. Making a model that captures the latent structure in continuous time is challenging and the quality of such models may be improved by incorporating the uncertainty of the latent information.

In this work, Raphaël introduces a method called Temporal Gaussian Network Embedding, which yields trajectories for entities in a low-dimensional space with explicit uncertainty. This yields a more accurate model that can be efficiently learned from data.

Curious to know more? Read it –open access– via the link below.