DLScanner : A parameter space scanner package assisted by deep learning methods

A. Hammad, Raymundo Ramos · 2024

Read DLScanner : A parameter space scanner package assisted by deep learning methods by A. Hammad, Raymundo Ramos (2024) 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-2412-19675.txt.

Description

Two physicists present a software package that uses deep learning to search the parameter spaces of physical theories, in this 2024 paper. Theories beyond the standard model come with many free numbers. Finding the combinations that match experiments means scanning a vast space, and older deep learning scanners converged slowly in many dimensions and generalized poorly from random samples.

Their package, DLScanner, attacks both faults. A similarity network first maps sampled points into a space where promising points cluster together and hopeless ones drift apart, which speeds convergence. A dynamic sampling rule, borrowed from an adaptive method called VEGAS, then proposes new points where they are needed most and improves the map as data accumulate. The authors report substantial gains in speed and accuracy over rival scanners. For model builders, the tool promises faster answers about which versions of a theory remain alive.

Catalog record

CREATORA. Hammad, Raymundo Ramos
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE79 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-19675

Browse the full library · About the Tower of Babel