A compact, interrogable Tree Tensor Network Jet Tagger on Lund declustering trees
Abstract
Jet taggers built on the Lund declustering tree encode QCD radiation structure in the inputs and process it with graph networks. We ask whether the tree itself can also constrain the computation, and what becomes accessible when it does. For this, we construct a novel quantum-inspired Tree Tensor Network (TTN), where a single tensor core is shared among the internal nodes of the Cambridge/Aachen declustering tree, employing Lund coordinates, and with no particle-identification or detector-level information. The construction is based on a strong prior on the radiation history rather than on the constituents, complementary to compact equivariant taggers. Because the same learned map acts identically at every node, the trained model can be evaluated on truncated trees and its latent space compared across recursions, and used to localise where on the Lund jet plane the Pythia and Herwig jet representations diverge, showing the regions of generator-dependent concentration. The χ=8 model has 4014 trainable parameters against approximately 391,000 for LundNet. We evaluate boosted top tagging and quark/gluon discrimination at particle level on Pythia and Herwig samples, in both native (Pythia→Pythia) and transfer (Pythia→Herwig) configurations. Despite differing in capacity by two orders of magnitude, the TTN reaches native performance close to LundNet in quark/gluon discrimination, slightly underperforming in top tagging, with a similar Pythia-to-Herwig transfer behavior in both tasks. Finally, we show the behavior of the network as a function of the bond dimension, and as a function of the perturbative scale.
Related
- Approximating Local Properties by Tensor Network States With Constant Bond Dimension
- Tensor network representations for intrinsically mixed-state topological orders
- Tensor-network approach to quantum optical state evolution beyond the Fock basis
- Is the Number of Foreshocks Underestimated? The Role of Declustering Symmetry
- CTMRG as a tensor-network message-passing algorithm
BibTeX
@misc{napolitano2026compact,
title = {{A compact, interrogable Tree Tensor Network Jet Tagger on Lund declustering trees}},
author = {Fabrizio Napolitano and Luca Della Penna and Tommaso Tedeschi and Livio Fanò},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.12408},
eprint = {2610.12408},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.12408},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - A compact, interrogable Tree Tensor Network Jet Tagger on Lund declustering trees
AU - Fabrizio Napolitano
AU - Luca Della Penna
AU - Tommaso Tedeschi
AU - Livio Fanò
PY - 2026
DA - 2026/10/08
PB - arXiv
DO - 10.48550/arxiv.2610.12408
UR - https://arxiv.org/abs/2610.12408
AB - Jet taggers built on the Lund declustering tree encode QCD radiation structure in the inputs and process it with graph networks. We ask whether the tree itself can also constrain the computation, and what becomes accessible when it does. For this, we construct a novel quantum-inspired Tree Tensor Network (TTN), where a single tensor core is shared among the internal nodes of the Cambridge/Aachen declustering tree, employing Lund coordinates, and with no particle-identification or detector-level information. The construction is based on a strong prior on the radiation history rather than on the constituents, complementary to compact equivariant taggers. Because the same learned map acts identically at every node, the trained model can be evaluated on truncated trees and its latent space compared across recursions, and used to localise where on the Lund jet plane the Pythia and Herwig jet representations diverge, showing the regions of generator-dependent concentration. The χ=8 model has 4014 trainable parameters against approximately 391,000 for LundNet. We evaluate boosted top tagging and quark/gluon discrimination at particle level on Pythia and Herwig samples, in both native (Pythia→Pythia) and transfer (Pythia→Herwig) configurations. Despite differing in capacity by two orders of magnitude, the TTN reaches native performance close to LundNet in quark/gluon discrimination, slightly underperforming in top tagging, with a similar Pythia-to-Herwig transfer behavior in both tasks. Finally, we show the behavior of the network as a function of the bond dimension, and as a function of the perturbative scale.
ER -