https://doi.org/10.1051/epjconf/202637001032
ARGUS: The AI Eye for CERN - Solving Extreme Computational Physics
1 Department of Computer Science Engineering and Artificial Intelligence & Machine Learning, Institute of Engineering and Management, Kolkata, West Bengal, India
2 Department of Computer Science and Business Systems, Institute of Engineering and Management, Kolkata, West Bengal, India
3 Department of Computer Science Engineering, Institute of Engineering and Management, Kolkata, West Bengal, India
* Corresponding author: n This email address is being protected from spambots. You need JavaScript enabled to view it.
Published online: 29 May 2026
Abstract
The present problem with High Luminosity Large Hadron Collider experiment is reconstructing particle trajectories from massive event data. When hadrons collide, the resulting particles form a point cloud, but only charged particles are relevant for further analysis and First Level Event Selection (FLES). In current detectors, sensors can only register whether a particle passed through them, making trajectory reconstruction difficult. Traditional Combinatorial Kalman Filters (CKF) try many possible track combinations, which causes a combinatorial explosion. Graph Neural Networks have also been applied, but their edge-heavy computation reduces speed. Vision Transformer (ViT)-based methods improved the task by treating it as a computer vision problem, but they still scale quadratically. We propose a new heuristic-based method that reconstructs trajectories in linear time, O(N). Our algorithm is about 23% faster than ViT and 140% faster than CKF, while maintaining strong trajectory estimation performance, providing a much faster and simpler alternative for realtime particle track reconstruction in HL-LHC experiments.
© The Authors, published by EDP Sciences, 2026
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
