Regression on high-dimensional phase space
Abstract
Machine learning tasks in particle physics are typically expressed in terms of regression, or relaxation toward some optimal solution as defined by an objective or loss function. In this paper, we introduce a large multiplicity expansion for analyzing regression tasks, which enables significant analytic understanding. For concreteness, we study regression with energy flow polynomials (EFPs), a complete linear basis of infrared and collinear safe observables on Lorentz invariant phase space. While the expectation value of any EFP scales like a constant in the large multiplicity limit, we prove that the covariance on uniform phase space of any two EFPs is highly suppressed and scales like the inverse of the squared multiplicity. This very unintuitive result relies on detailed cancellation guaranteed by the interplay of permutation invariance, momentum conservation, and rotational invariance. This scaling of the covariance with multiplicity further implies that the coefficients in an expansion of EFPs must scale quadratically in multiplicity. We present an example of regression with EFPs as applied to e+e-→ hadrons events, and show how the regression can be solved analytically through the complementary expansions of the large multiplicity limit and of familiar fixed-order perturbation theory of QCD.
Related
- Nonlinearity-induced Topological Phase Transitions in One-dimensional Systems with Chiral Symmetry: A Real-Space Treatment
- Exact Solution of the (2+1)-Dimensional Dunkl-Klein-Gordon Oscillator with a Uniform Magnetic Field in Non-Commutative Phase Space
- Three-Dimensional Coulomb Discrete Spectrum via Symplectic Geometry and Phase-Space Constraints: A Geometric Framework for Condensed-Matter Coulomb-Like Systems
- Statistical Parametric Mapping reveals sex-specific, phase-dependent differences in bilateral three-dimensional ground reaction force during Wushu Changbing PiJi in high-level athletes
- Rough-target ISAL autofocus under coupled multi-source phase disturbances: coherence-weighted circular regression across multiple range lines
BibTeX
@misc{larkoski2026regression,
title = {{Regression on high-dimensional phase space}},
author = {Andrew J. Larkoski},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.36073},
eprint = {2610.36073},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.36073},
}
RIS
TY - UNPB
TI - Regression on high-dimensional phase space
AU - Andrew J. Larkoski
PY - 2026
DA - 2026/10/26
PB - arXiv
DO - 10.48550/arxiv.2610.36073
UR - https://arxiv.org/abs/2610.36073
AB - Machine learning tasks in particle physics are typically expressed in terms of regression, or relaxation toward some optimal solution as defined by an objective or loss function. In this paper, we introduce a large multiplicity expansion for analyzing regression tasks, which enables significant analytic understanding. For concreteness, we study regression with energy flow polynomials (EFPs), a complete linear basis of infrared and collinear safe observables on Lorentz invariant phase space. While the expectation value of any EFP scales like a constant in the large multiplicity limit, we prove that the covariance on uniform phase space of any two EFPs is highly suppressed and scales like the inverse of the squared multiplicity. This very unintuitive result relies on detailed cancellation guaranteed by the interplay of permutation invariance, momentum conservation, and rotational invariance. This scaling of the covariance with multiplicity further implies that the coefficients in an expansion of EFPs must scale quadratically in multiplicity. We present an example of regression with EFPs as applied to e^+e^-→ hadrons events, and show how the regression can be solved analytically through the complementary expansions of the large multiplicity limit and of familiar fixed-order perturbation theory of QCD.
ER -