A Language Model for Particle Tracking

TL;DR

This paper introduces TrackingBERT, a novel language model for particle tracking at the LHC, enabling better generalization and multi-task capabilities through a tokenized detector representation.

Contribution

The paper presents a new tokenized detector representation and trains a BERT-based model, TrackingBERT, for particle tracking, pioneering a foundational model approach in this domain.

Findings

TrackingBERT achieves effective particle tracking performance.

The model provides latent detector embeddings for auxiliary tasks.

First application of language models to particle detector data.

Abstract

Particle tracking is crucial for almost all physics analysis programs at the Large Hadron Collider. Deep learning models are pervasively used in particle tracking related tasks. However, the current practice is to design and train one deep learning model for one task with supervised learning techniques. The trained models work well for tasks they are trained on but show no or little generalization capabilities. We propose to unify these models with a language model. In this paper, we present a tokenized detector representation that allows us to train a BERT model for particle tracking. The trained BERT model, namely TrackingBERT, offers latent detector module embedding that can be used for other tasks. This work represents the first step towards developing a foundational model for particle detector understanding.