11
0

Structure Learning in Graphical Models from Indirect Observations

Hang Zhang
Afshin Abdi
Faramarz Fekri
Abstract

This paper considers learning of the graphical structure of a pp-dimensional random vector XRpX \in R^p using both parametric and non-parametric methods. Unlike the previous works which observe xx directly, we consider the indirect observation scenario in which samples yy are collected via a sensing matrix ARd×pA \in R^{d\times p}, and corrupted with some additive noise ww, i.e, Y=AX+WY = AX + W. For the parametric method, we assume XX to be Gaussian, i.e., xRpN(μ,Σ)x\in R^p\sim N(\mu, \Sigma) and ΣRp×p\Sigma \in R^{p\times p}. For the first time, we show that the correct graphical structure can be correctly recovered under the indefinite sensing system (d<pd < p) using insufficient samples (n<pn < p). In particular, we show that for the exact recovery, we require dimension d=Ω(p0.8)d = \Omega(p^{0.8}) and sample number n=Ω(p0.8log3p)n = \Omega(p^{0.8}\log^3 p). For the nonparametric method, we assume a nonparanormal distribution for XX rather than Gaussian. Under mild conditions, we show that our graph-structure estimator can obtain the correct structure. We derive the minimum sample number nn and dimension dd as n(deg)4log4nn\gtrsim (deg)^4 \log^4 n and dp+(deglog(dp))β/4d \gtrsim p + (deg\cdot\log(d-p))^{\beta/4}, respectively, where deg is the maximum Markov blanket in the graphical model and β>0\beta > 0 is some fixed positive constant. Additionally, we obtain a non-asymptotic uniform bound on the estimation error of the CDF of XX from indirect observations with inexact knowledge of the noise distribution. To the best of our knowledge, this bound is derived for the first time and may serve as an independent interest. Numerical experiments on both real-world and synthetic data are provided confirm the theoretical results.

View on arXiv
Comments on this paper