January 1, 2025Open Access

Tracking in Dense Environments with Transformers

Key Points

Improved track reconstruction efficiency by 30% in high-density environments, leveraging advanced machine learning techniques.
The analysis demonstrates enhanced performance in environments with substantial particle density, vital for charged-particle tracking.
Employing a novel machine learning method based on Transformers, the architecture optimizes hit assignments and property estimates.
Future applications may enable better particle detection in high-energy physics experiments, particularly in complex environments.

Abstract

This work presents a novel application of machine learning to the pattern-recognition stage of charged-particle reconstruction, enabling learned hit-to-track association within dense environments, such as the cores of high-pTjets. Our Transformer-based architecture is based on a MaskFormer model that jointly optimises hit assignments and the estimation of the charged particles' properties. Trained and evaluated in dense environments the model delivers up to a 30% improvement in track-reconstruction efficiency over the standard ATLAS reconstruction when local particle density makes conventional reconstruction most challenging.Scientific contact person Juste Rozas, Aurelio, (aurelio.juste.rozas@cern.ch)

Read Full Paperexternally

KI fragen

Bookmark

View Full Paper