What exactly did the Transformer learn from our physics data?

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Erdmann, Martin, Langner, Niklas, Schulte, Josina, Wirtz, Dominik
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917402236682240
author Erdmann, Martin
Langner, Niklas
Schulte, Josina
Wirtz, Dominik
author_facet Erdmann, Martin
Langner, Niklas
Schulte, Josina
Wirtz, Dominik
contents Transformer networks excel in scientific applications. We explore two scenarios in ultra-high-energy cosmic ray simulations to examine what these network architectures learn. First, we investigate the trained positional encodings in air showers which are azimuthally symmetric. Second, we visualize the attention values assigned to cosmic particles originating from a galaxy catalog. In both cases, the Transformers learn plausible, physically meaningful features.
format Preprint
id arxiv_https___arxiv_org_abs_2505_21042
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle What exactly did the Transformer learn from our physics data?
Erdmann, Martin
Langner, Niklas
Schulte, Josina
Wirtz, Dominik
Instrumentation and Methods for Astrophysics
High Energy Physics - Experiment
Transformer networks excel in scientific applications. We explore two scenarios in ultra-high-energy cosmic ray simulations to examine what these network architectures learn. First, we investigate the trained positional encodings in air showers which are azimuthally symmetric. Second, we visualize the attention values assigned to cosmic particles originating from a galaxy catalog. In both cases, the Transformers learn plausible, physically meaningful features.
title What exactly did the Transformer learn from our physics data?
topic Instrumentation and Methods for Astrophysics
High Energy Physics - Experiment
url https://arxiv.org/abs/2505.21042