arXiv subjects

Partha Sarkar

Publications and source records attributed to Partha Sarkar.

4 recordsLinked to original sources

arXiv2026

Bayesian Graphical Models under Positivity Constraints: A Scalable generalized likelihood Approach

We develop a computationally scalable Bayesian framework for precision matrix estimation in Gaussian graphical models under total positivity constraints. To overcome the high computational cost of the Gaussian likelihood, we adopt a generalized Bayesian approach based on the $D$-trace loss, which eliminates the log-determinant term and enables efficient optimization while allowing relaxation of positive definiteness during sampling. Sparsity is induced via spike-and-slab priors, and the resulting generalized posterior is shown to be proper under mild conditions. Our primary contribution is a suite of efficient posterior sampling algorithms tailored to high-dimensional settings. Starting from a component-wise Gibbs sampler, we introduce a novel data augmentation scheme that induces conditional independence among precision matrix entries, enabling joint updates. By exploiting the Gram structure of the sample covariance matrix, we further develop a fast matrix-normal sampler that significantly reduces per-iteration complexity in high-dimensional settings. An interweaving strategy combines augmented and direct updates to improve mixing without sacrificing scalability. Experiments on synthetic and financial data demonstrate substantial computational gains over existing methods, while maintaining competitive estimation accuracy and improved recovery of structured dependencies.

stat.ME↗

arXiv2026

Moment bounds for condition numbers and singular values of high-dimensional Gaussian random matrices: Applications and limitations

Spectral properties of Gram matrices are central to high dimensional asymptotic analyses of statistical estimators in regression and covariance estimation. These properties, in turn, depend critically on the extreme singular values and condition numbers of Gaussian random matrices. For many applications, sharp positive and negative moment bounds for these quantities are required to control expected prediction risk and related performance metrics. Although extensive work provides concentration and tail bounds for extreme singular values of Gaussian random matrices, these results do not readily yield the moment bounds needed in such analyses. Motivated by this gap, we establish non asymptotic moment bounds for arbitrary positive moments of the largest singular value and arbitrary negative moments of the smallest singular value, and uniform bounds for arbitrary positive moments of the condition number of high dimensional Gaussian random matrices. We demonstrate the utility of these bounds by applying them to derive explicit risk guarantees in high dimensional regression and covariance estimation, as well as to obtain bounds on the mean iteration complexity of gradient descent for solving Gram linear systems. Finally, we present counterexamples demonstrating that the positive condition number moment bounds and negative smallest singular value moment bounds cannot, in general, be extended to the broader class of sub Gaussian random matrices.

math.ST↗

arXiv2023

High-Dimensional Bernstein Von-Mises Theorems for Covariance and Precision Matrices

This paper aims to examine the characteristics of the posterior distribution of covariance/precision matrices in a "large $p$, large $n$" scenario, where $p$ represents the number of variables and $n$ is the sample size. Our analysis focuses on establishing asymptotic normality of the posterior distribution of the entire covariance/precision matrices under specific growth restrictions on $p_n$ and other mild assumptions. In particular, the limiting distribution turns out to be a symmetric matrix variate normal distribution whose parameters depend on the maximum likelihood estimate. Our results hold for a wide class of prior distributions which includes standard choices used by practitioners. Next, we consider Gaussian graphical models which induce sparsity in the precision matrix. Asymptotic normality of the corresponding posterior distribution is established under mild assumptions on the prior and true data-generating mechanism.

math.ST↗

arXiv2023

Posterior consistency in multi-response regression models with non-informative priors for the error covariance matrix in growing dimensions

The Inverse-Wishart (IW) distribution is a standard and popular choice of priors for covariance matrices and has attractive properties such as conditional conjugacy. However, the IW family of priors has crucial drawbacks, including the lack of effective choices for non-informative priors. Several classes of priors for covariance matrices that alleviate these drawbacks, while preserving computational tractability, have been proposed in the literature. These priors can be obtained through appropriate scale mixtures of IW priors. However, in the era of increasing dimensionality, the posterior consistency of models that incorporate such priors has not been investigated. We address this issue for the multi-response regression setting ($q$ responses, $n$ samples) under a wide variety of IW scale mixture priors for the error covariance matrix. Posterior consistency and contraction rates for both the regression coefficient matrix and the error covariance matrix are established in the ``large $q$, large $n$'' setting under mild assumptions on the true data-generating covariance matrix and relevant hyperparameters. In particular, the number of responses $q_n$ is allowed to grow with $n$, but with $q_n = o(n)$. Also, some results related to the inconsistency of the posterior distribution and posterior mean for $q_n/n \to \gamma$, where $\gamma \in (0,\infty)$ are provided.

math.ST↗