Abstract:We study the problem of change point detection for covariance matrices in high dimensions. We assume that we observe a sequence {X_i}_{i=1,...,n} of independent and centered p-dimensional sub-Gaussian random vectors whose covariance matrices are piecewise constant. Our task is to recover with high accuracy the number and locations of the change points, which are assumed unknown. Our generic model setting allows for all the model parameters to change with n, including the dimension p, the minimal spacing between consecutive change points, the magnitude of smallest change size and the maximal Orlicz- 2 norm of the covariance matrices of the sample points. Without assuming any additional structural assumption, such as low rank matrices or having sparse principle components, we set up a general framework and a benchmark result for the covariance change point detection problem. We introduce two procedures, one based on the binary segmentation algorithm (e.g. Vostrikova, 1981) and the other on its extension known as wild binary segmentation of Fryzlewicz (2014), and demonstrate that, under suitable conditions, both procedures are able to consistently es- timate the number and locations of change points. Our second algorithm, called Wild Binary Segmentation through Independent Projection (WBSIP), is shown to be optimal in the sense of allowing for the minimax scaling in all the relevant parameters. Our minimax analysis reveals a phase transition effect based on the problem of change point localization. To the best of our knowledge, this type of results has not been established elsewhere in the high-dimensional change point detection literature.

Covariance alignment: from maximum likelihood estimation to Gromov-Wasserstein

Neural Entropic Gromov-Wasserstein Alignment

Robust and High-Order Correlation Alignment for Unsupervised Domain Adaptation

Scalable unsupervised alignment of general metric and non-metric structures

Gaussian Database Alignment and Gaussian Planted Matching

Improved Covariance Modeling For Maxim Unilikelihood Multiple Subspace Transformations

Database Alignment with Gaussian Features

Revisiting invariances and introducing priors in Gromov-Wasserstein distances

Hierarchical Optimal Transport for Multimodal Distribution Alignment

On Robust Cross Domain Alignment

A Variance Minimization Criterion to Feature Selection Using Laplacian Regularization

Distributional Preference Alignment of LLMs via Optimal Transport

Entropic Gromov-Wasserstein Distances: Stability and Algorithms

Limit Laws for Gromov-Wasserstein Alignment with Applications to Testing Graph Isomorphisms

Covariance Model with General Linear Structure and Divergent Parameters

Aligning Embeddings and Geometric Random Graphs: Informational Results and Computational Approaches for the Procrustes-Wasserstein Problem

Alignment and Correspondence Using Singular Value Decomposition

A Data-dependent Approach for High Dimensional (Robust) Wasserstein Alignment

Improved covariance modeling for maximum likelihood multiple subspace transformations [speech recognition applications]

Optimal Covariance Change Point Localization in High Dimension

Fused Gromov-Wasserstein Alignment for Hawkes Processes