This is SParTA: A modular, efficient and scalable transformer architecture for jet-tagging
Abstract
We present the Sparse Particle-interaction Transformer Architecture (SParTA), a modular, scalable, and efficient transformer for jet tagging at high-energy colliders. Exploiting the steeply falling transverse-momentum (pT) spectrum of jet constituents, SParTA factorizes constituent processing into two topological regimes. Ordered by descending pT, only the N hardest constituents (the hard core) participate in quadratic O(N2) differential interaction attention driven by pairwise physical and learned relational features. The remaining soft constituents are integrated via linear O(lN) cross-attention to the hard core, preserving soft-radiation and color-flow information while avoiding the O(l2) memory bottleneck of full self-attention over all l constituents. Pairwise interaction features are computed dynamically in-network from constituent four-momenta and conditioned on global jet-level substructure observables. On the quark-gluon tagging benchmark, a 300K-parameter SParTA model achieves an AUC of 92.02% and accuracy of 85.06%, matching the 2.14M-parameter Particle Transformer (ParT) at one-seventh the parameter count and outperforming IAFormer by 21% in background rejection at 30% signal efficiency. On the boosted top-tagging benchmark, SParTA outperforms ParT and ParticleNet with 511K parameters, while we analyze the architectural trade-offs relative to fully dense multi-layer interaction networks.
Related
- PolyUAnalog: A Modular and Scalable Open-Source Architecture for Polyphonic Analog Synthesis
- Scalable, Energy-Efficient Optical-Neural Architecture for Multiplexed Deepfake Video Detection
- QGrAT: Quantum Grade-Aware Transformer for Jet Tagging
- How Far Can Foundation Models Be Distilled for Efficient Jet Tagging?
- Continuous vision transformer-based neural operator for scalable beam propagation modeling in optical fibers
BibTeX
@misc{baruah2026this,
title = {{This is SParTA: A modular, efficient and scalable transformer architecture for jet-tagging}},
author = {Rajneil Baruah and Subhadeep Mondal and Sunando Kumar Patra and Satyajit Roy},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.27106},
eprint = {2610.27106},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.27106},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - This is SParTA: A modular, efficient and scalable transformer architecture for jet-tagging
AU - Rajneil Baruah
AU - Subhadeep Mondal
AU - Sunando Kumar Patra
AU - Satyajit Roy
PY - 2026
DA - 2026/10/19
PB - arXiv
DO - 10.48550/arxiv.2610.27106
UR - https://arxiv.org/abs/2610.27106
AB - We present the Sparse Particle-interaction Transformer Architecture (SParTA), a modular, scalable, and efficient transformer for jet tagging at high-energy colliders. Exploiting the steeply falling transverse-momentum (p_T) spectrum of jet constituents, SParTA factorizes constituent processing into two topological regimes. Ordered by descending p_T, only the N hardest constituents (the hard core) participate in quadratic O(N^2) differential interaction attention driven by pairwise physical and learned relational features. The remaining soft constituents are integrated via linear O(lN) cross-attention to the hard core, preserving soft-radiation and color-flow information while avoiding the O(l^2) memory bottleneck of full self-attention over all l constituents. Pairwise interaction features are computed dynamically in-network from constituent four-momenta and conditioned on global jet-level substructure observables. On the quark-gluon tagging benchmark, a 300K-parameter SParTA model achieves an AUC of 92.02% and accuracy of 85.06%, matching the 2.14M-parameter Particle Transformer (ParT) at one-seventh the parameter count and outperforming IAFormer by 21% in background rejection at 30% signal efficiency. On the boosted top-tagging benchmark, SParTA outperforms ParT and ParticleNet with 511K parameters, while we analyze the architectural trade-offs relative to fully dense multi-layer interaction networks.
ER -