Joint Independence Testing for Latent Position Random Graphs
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≥ 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 (n−1/2), 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≥ 2 LPRGs, (ii) we prove that it can be estimated at a rate Õ(n−1/3) 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.
Related
- On the Stability of the Independence Number in Random Distance Graphs
- Two-Sample Testing for Inhomogeneous Random Graphs in Non-Integral Lr Norms
- Two-Sample Testing for Random Graphs without Vertex Correspondence
- Tree-independence number VIII. 3PC-free graphs have logarithmic tree-independence number
- Strong complete immersions in graphs with independence number two
BibTeX
@misc{wen2026joint,
title = {{Joint Independence Testing for Latent Position Random Graphs}},
author = {Xiaoyi Wen and Zoltan Szabo},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.34544},
eprint = {2610.34544},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.34544},
}
RIS
TY - UNPB
TI - Joint Independence Testing for Latent Position Random Graphs
AU - Xiaoyi Wen
AU - Zoltan Szabo
PY - 2026
DA - 2026/10/25
PB - arXiv
DO - 10.48550/arxiv.2610.34544
UR - https://arxiv.org/abs/2610.34544
AB - 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≥ 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 (n^−1/2), 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≥ 2 LPRGs, (ii) we prove that it can be estimated at a rate Õ(n^−1/3) 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.
ER -