The Poisson tensor completion parametric estimator

Dunlavy, Daniel Michael, Lehoucq, Richard B., Mayer, Carolyn Doo et al. · 2026

Read The Poisson tensor completion parametric estimator by Dunlavy, Daniel Michael, Lehoucq, Richard B., Mayer, Carolyn Doo et al. (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/doe-osti-3023847-the-poisson-tensor-completion-parametric.txt.

Description

Daniel M Dunlavy, Richard B Lehoucq, Carolyn D Mayer, and Arvind Prasadan of Sandia National Laboratories introduce a new statistical estimator in SAND2026-18285R. Counting samples of a multi-variable distribution into histogram bins leaves many bins nearly empty, which makes plain histograms shaky. Their Poisson tensor completion estimator fills those sparse bins by exploiting links between samples.

The key move is to treat histogram bins as a space partitioning of counts from a non-homogeneous Poisson process. A low-rank Poisson tensor decomposition then completes the mean measure over all bins, even those holding few or no samples, and the estimates come out non-negative by construction. For sub-Gaussian distributions, the authors show a substantial gain over standard histogram estimators, resting on the concentration of norm phenomenon.

Catalog record

CREATORDunlavy, Daniel Michael, Lehoucq, Richard B., Mayer, Carolyn Doo et al.
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE64 KB
LICENSEPublic Domain
RIGHTSPUBLIC DOMAIN
SOURCEDOE OSTI

Source: https://www.osti.gov/biblio/3023847

Browse the full library · About the Tower of Babel