Efficient ML-Assisted Particle Track Reconstruction Designs

¹ High-Energy Physics, Radboud University, Nijmegen, The Netherlands
² National Institute for Subatomic Physics (Nikhef), Amsterdam, The Netherlands
³ Institute for Computing and Information Sciences, Radboud University, Nijmegen, The Netherlands
⁴ Intelligent Data Analysis Laboratory (IDAL), Department of Electronic Engineering, ETSE-UV, University of Valencia, Valencia, Spain
⁵ Valencian Graduate School and Research Network of Artificial Intelligence (ValgrAI), Valencia, Spain
⁶ Faculty of Engineering Technology, University of Twente, Enschede, The Netherlands
⁷ Instituto de Física Corpuscular, University of Valencia, Valencia, Spain
⁸ Institute of Physics, University of Amsterdam, Amsterdam, The Netherlands
⁹ SURF, Amsterdam, The Netherlands

^* e-mail: scaron@nikhef.nl
^** e-mail: nadezhda.dobreva@ru.nl
^*** e-mail: antonio.ferrer-sanchez@uv.es
^**** e-mail: jose.d.martin@uv.es
^† e-mail: uodyurt@nikhef.nl
^‡ e-mail: rruiz@ific.uv.es
^§ e-mail: zwolffs@uva.nl
^¶ e-mail: yue.zhao@surf.nl

Published online: 7 October 2025

Abstract

Track reconstruction is a crucial part of High Energy Physics experiments. Traditional methods for the task, relying on Kalman Filters, scale poorly with detector occupancy. In the context of the upcoming High Luminosity-LHC, solutions based on Machine Learning (ML) and deep learning are very appealing. We investigate the feasibility of training multiple ML architectures to infer track-defining parameters from detector signals, for the application of offline reconstruction. We study and compare three Transformer model designs, as well as a U-Net architecture. We describe in detail the two most promising approaches and benchmark the pipelines for physics performance and inference speed on methodically simplified datasets, generated by the recently developed simulation framework, REDuced VIrtual Detector (REDVID). Our second batch of simplified datasets are derived from the TrackML dataset. Our preliminary results show promise for the application of such deep learning techniques on more realistic data for tracking, as well as efficient elimination of solutions.