Learning sparse linear dynamic networks in a hyper-parameter free setting

TL;DR

This paper introduces a hyperparameter-free, efficient iterative method called SPICE for estimating the topology and dynamics of sparse linear networks, applicable under various noise conditions without prior knowledge.

Contribution

The paper presents a novel hyperparameter-free approach for network topology and dynamics estimation using SPICE, which is computationally efficient and does not require prior assumptions.

Findings

Effective in estimating network topology and dynamics

Works under varying noise levels across modules

Applicable to directed and undirected networks

Abstract

We address the issue of estimating the topology and dynamics of sparse linear dynamic networks in a hyperparameter-free setting. We propose a method to estimate the network dynamics in a computationally efficient and parameter tuning-free iterative framework known as SPICE (Sparse Iterative Covariance Estimation). The estimated dynamics directly reveal the underlying topology. Our approach does not assume that the network is undirected and is applicable even with varying noise levels across the modules of the network. We also do not assume any explicit prior knowledge on the network dynamics. Numerical experiments with realistic dynamic networks illustrate the usefulness of our method.