Mathematics Colloquia and Seminars

Return to Colloquia & Seminar listing

A common task in data science is to learn a graph representation from real-life data, which can be utilized for decision making such as finding the key influencers in social networks. This talk presents recent results on graph learning using low pass graph signal processing (GSP) models. In other words, the purported data generative model consists of a graph filter which only retains the low frequencies content, i.e., effectively implying the common “smooth graph signals”. We motivate by showing the prevalence of low pass GSP models in dynamics for social networks, climate networks, etc. We then demonstrate how the spectral properties of low pass graph signals can be leveraged for graph topology learning from nodal observation. We will discuss recent results that are motivated from low pass models, including (i) graph topology learning from partial observation, (ii) multiplex graph learning, and (iii) detection of low pass graph signals, etc. Other applications of low pass GSP will also be discussed.