Neural Scaling Laws for Jet Generation

Oz Amram, Darius A. Faroughy, Tjarko Gerdes, Anna Hallin, Gregor Kasieczka, Michael Krämer, Humberto Reyes-Gonzalez, David Shih · 2026

Read Neural Scaling Laws for Jet Generation by Oz Amram, Darius A. Faroughy, Tjarko Gerdes, Anna Hallin, Gregor Kasieczka, Michael Krämer, Humberto Reyes-Gonzalez, David Shih (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-2605-28940.txt.

Description

Oz Amram and colleagues test whether neural scaling laws, familiar from language models, also govern generative models of particle jets. They recover the expected logarithmic model-size scaling for next-token prediction loss. A sliced Wasserstein distance over five physical quantities tracks that loss closely, making the loss a sound proxy for physics quality. Dataset-size and compute scaling prove much weaker. The authors explain this with a learnable window: jet constituents saturate quickly, unlike language tokens, partly because QCD radiation is random.

Catalog record

CREATOROz Amram, Darius A. Faroughy, Tjarko Gerdes, Anna Hallin, Gregor Kasieczka, Michael Krämer, Humberto Reyes-Gonzalez, David Shih
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE51237
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.28940

Browse the full library · About the Tower of Babel