35
5

Learning the Structure of Large Networked Systems Obeying Conservation Laws

Abstract

Many networked systems such as electric networks, the brain, and social networks of opinion dynamics are known to obey conservation laws. Examples of this phenomenon include the Kirchoff laws in electric networks and opinion consensus in social networks. Conservation laws in networked systems may be modeled as balance equations of the form X=BYX = B^{*} Y, where the sparsity pattern of BB^{*} captures the connectivity of the network, and Y,XRpY, X \in \mathbb{R}^p are vectors of "potentials" and "injected flows" at the nodes respectively. The node potentials YY cause flows across edges and the flows XX injected at the nodes are extraneous to the network dynamics. In several practical systems, the network structure is often unknown and needs to be estimated from data. Towards this, one has access to samples of the node potentials YY, but only the statistics of the node injections XX. Motivated by this important problem, we study the estimation of the sparsity structure of the matrix BB^{*} from nn samples of YY under the assumption that the node injections XX follow a Gaussian distribution with a known covariance ΣX\Sigma_X. We propose a new 1\ell_{1}-regularized maximum likelihood estimator for this problem in the high-dimensional regime where the size of the network pp is larger than sample size nn. We show that this optimization problem is convex in the objective and admits a unique solution. Under a new mutual incoherence condition, we establish sufficient conditions on the triple (n,p,d)(n,p,d) for which exact sparsity recovery of BB^{*} is possible with high probability; dd is the degree of the graph. We also establish guarantees for the recovery of BB^{*} in the element-wise maximum, Frobenius, and operator norms. Finally, we complement these theoretical results with experimental validation of the performance of the proposed estimator on synthetic and real-world data.

View on arXiv
Comments on this paper