Transformers for Charged Particle Track Reconstruction in High-Energy Physics
Bibliographic record
Abstract
Charged particle reconstruction, the identification and characterization of particles from collision data, is fundamental to nearly all research at particle colliders like the Large Hadron Collider (LHC). With the High-Luminosity upgrade (HL-LHC), particle multiplicities will increase substantially, overwhelming traditional track reconstruction algorithms and presenting computational bottlenecks. Here, we introduce a proof of concept for a powerful new method for charged particle reconstruction inspired by state-of-the-art machine learning (ML) approaches in computer vision. Our model leverages transformer neural networks to efficiently filter relevant signals and fully reconstruct particle trajectories, directly tackling the computational complexity that traditional methods face. Evaluated on the widely used TrackML dataset, our approach achieves state-of-the-art tracking efficiency (97%) and a low fake rate (0.7%), requiring just 97 ms to reconstruct on average 1300 particle trajectories from 55,000 detector hits for particles with transverse momentum above 750 MeV. These results represent a significant milestone in both performance and speed, demonstrating a shift toward unified, scalable ML solutions that offer substantial improvements for collider experiments.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.006 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.005 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".