GLOW: A Unified Particle Flow Transformer

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kobylianskii, Dmitrii, Van Stroud, Samuel, Wong, Kwok Yiu, Hart, Max, Dreyer, Etienne, Gross, Eilam, Facini, Gabriel, Scanlon, Tim
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915466625155072
author Kobylianskii, Dmitrii
Van Stroud, Samuel
Wong, Kwok Yiu
Hart, Max
Dreyer, Etienne
Gross, Eilam
Facini, Gabriel
Scanlon, Tim
author_facet Kobylianskii, Dmitrii
Van Stroud, Samuel
Wong, Kwok Yiu
Hart, Max
Dreyer, Etienne
Gross, Eilam
Facini, Gabriel
Scanlon, Tim
contents We present GLOW, a transformer-based particle flow model that combines incidence matrix supervision from HGPflow with a MaskFormer architecture. Evaluated on CLIC detector simulations, GLOW achieves state-of-the-art performance and, together with prior work, demonstrates that a single unified transformer architecture can effectively address diverse reconstruction tasks in particle physics.
format Preprint
id arxiv_https___arxiv_org_abs_2508_20092
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GLOW: A Unified Particle Flow Transformer
Kobylianskii, Dmitrii
Van Stroud, Samuel
Wong, Kwok Yiu
Hart, Max
Dreyer, Etienne
Gross, Eilam
Facini, Gabriel
Scanlon, Tim
High Energy Physics - Experiment
High Energy Physics - Phenomenology
We present GLOW, a transformer-based particle flow model that combines incidence matrix supervision from HGPflow with a MaskFormer architecture. Evaluated on CLIC detector simulations, GLOW achieves state-of-the-art performance and, together with prior work, demonstrates that a single unified transformer architecture can effectively address diverse reconstruction tasks in particle physics.
title GLOW: A Unified Particle Flow Transformer
topic High Energy Physics - Experiment
High Energy Physics - Phenomenology
url https://arxiv.org/abs/2508.20092