Leveraging sparsity to improve no-U-turn sampling efficiency for hierarchical Bayesian models

Cole C. Monnahan, Kasper Kristensen, James T. Thorson, Bob Carpenter · 2026

Read Leveraging sparsity to improve no-U-turn sampling efficiency for hierarchical Bayesian models by Cole C. Monnahan, Kasper Kristensen, James T. Thorson, Bob Carpenter (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-02437.txt.

Description

Cole Monnahan, Kasper Kristensen, James Thorson, and Bob Carpenter speed up Bayesian sampling in this 2026 paper. Hierarchical models are workhorses for understanding natural processes. The no-U-turn sampler explores them well but slows under strong correlations. High dimensions make that slowdown worse and cap practical model complexity.

Their Sparse NUTS reshapes the posterior before sampling, using a sparse precision matrix. That matrix comes cheaply from an approximation at the posterior's peak. The reshaped problem then goes to standard software for sampling. Across seventeen varied case studies, sampling converged ten to a hundred times faster. Gains vanish for funnel shapes, wide tails, or multiple modes, they note. Huge models with sparse, correlated structure benefit most.

Catalog record

CREATORCole C. Monnahan, Kasper Kristensen, James T. Thorson, Bob Carpenter
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE70354
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.02437

Browse the full library · About the Tower of Babel