Measuring independence in complex network models with new statistical method

Joint Independence Testing for Latent Position Random Graphs

Machine LearningSocial and Information Networks

Summary

Graphs are used to represent relationships between things, like neurons in the brain or social connections. Sometimes these graphs are modeled by hidden features that affect who connects with whom, but these features are hard to observe directly. The authors developed a new way to check if multiple such graphs are independent or linked, even when these hidden features are only indirectly visible. Their method uses special math tools called kernels and spectral analysis, and it works faster and with fewer assumptions than previous methods. They showed it can be used for example on brain data to understand how different parts communicate during language processing.

What this means in practice

  • For network analysts: Evaluate whether multiple complex networks modeled by latent features operate independently or influence each other, improving analysis of interconnected systems.
  • For neuroscience data teams: Analyze brain network data to detect functional independence or interaction between subnetworks during cognitive tasks using a statistically robust method.

Authors

Xiaoyi Wen, Zoltan Szabo

Abstract

Kernel techniques are among the most powerful methods in data science and statistics. One of the key properties of kernels contributing to their numerous successful applications is that the associated reproducing kernel Hilbert space is expressive enough to encode probability distributions without loss of information or to capture the joint independence of $M\ge 2$ random variables, under mild conditions. Computationally, these measures (i) rely on the so-called Gram matrix which contains the pairwise similarity of sample points, and (ii) converge at a rate $O \left(n^{-\frac{1}{2}}\right)$, where $n$ denotes the sample size. We focus on measuring the independence of latent position random graphs (LPRGs), a model widely-applied for instance in neuroscience, sociology, and epidemiology. In LPRGs a latent representation is associated to each node. The similarity of these representations (meant via a kernel) determines whether two nodes are connected, making the Gram matrices only implicitly observable and the estimation challenging. Despite the fundamental importance of quantifying the independence of $M$ LPRGs, very little is known about the solution of this problem; existing techniques are restricted to $M=2$ components and only guarantee consistency without rates. In this work, (i) we define an independence measure called graph Hilbert-Schmidt independence criterion (GHSIC) for $M\ge 2$ LPRGs, (ii) we prove that it can be estimated at a rate $\tilde{O}\left(n^{-\frac{1}{3}}\right)$ relying on spectral decomposition of adjacency matrices, (iii) without the need to choose kernels and their hyperparameters. Numerical experiments demonstrate the effectiveness of the proposed GHSIC estimator both on synthetic examples and in the analysis of brain functional connectivity subnetworks in natural language perception.