High-Dimensional Bayesian Calibration of Expensive Nuclear Models with Differentiable Emulation

Jin Lei · 2026

Read High-Dimensional Bayesian Calibration of Expensive Nuclear Models with Differentiable Emulation by Jin Lei (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-2605-30980.txt.

Description

Lei offers DREAM, a strategy for Bayesian calibration of costly nuclear models. Missing exact gradients in legacy codes has blocked full fitting. DREAM samples the system offline, shrinks it by singular value factors, and rebuilds it online. Gradients then follow at the cost of one extra run per Monte Carlo step. Applied to coupled-channels deuteron-nickel elastic scattering with eighteen optical values, sampling settles on a single graphics chip in under ten minutes, yielding a posterior.

Catalog record

CREATORJin Lei
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE45209
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.30980

Browse the full library · About the Tower of Babel