arxiv: 2602.22776 · v3 · submitted 2026-02-26 · 🪐 quant-ph · hep-ex· physics.data-an

Recognition: no theorem link

Optimization-based Unfolding in High-Energy Physics

Simone Gasperini , Gianluca Bianco , Marco Lorusso , Carla Rieger , Michele Grossi

Authors on Pith no claims yet

Pith reviewed 2026-05-15 19:22 UTC · model grok-4.3

classification 🪐 quant-ph hep-exphysics.data-an

keywords unfoldingQUBOquantum annealinghigh-energy physicsregularizationdetector responseoptimizationhybrid solvers

0 comments

The pith

Reformulating unfolding as a regularized quadratic optimization problem enables the use of quantum annealing and hybrid solvers in high-energy physics.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper reformulates the unfolding problem in high-energy physics as a regularized quadratic optimization task. This is then encoded as a Quadratic Unconstrained Binary Optimization (QUBO) problem for direct application on quantum annealers and hybrid quantum-classical solvers. An open-source package QUnfold implements the method using both classical mixed-integer solvers and D-Wave's hybrid solver. Benchmarks on synthetic datasets with controlled distortions show competitive reconstruction accuracy compared to standard techniques like matrix inversion and iterative Bayesian unfolding. The approach naturally incorporates regularization into the objective function and offers a unified optimization framework for the task.

Core claim

Unfolding is recast as a quadratic optimization problem whose QUBO encoding can be solved by quantum-compatible methods to recover the true distribution from detector-level data with accuracy matching established classical techniques.

What carries the argument

The QUBO representation derived from the regularized quadratic objective, which allows implementation on quantum annealing and hybrid solvers while embedding regularization directly.

If this is right

The method accommodates regularization naturally within the optimization objective rather than as a post-processing step.
Competitive accuracy is achieved across multiple test distributions using synthetic data with known distortions.
A practical software package integrates classical and quantum solvers for unfolding tasks.
The framework provides a pathway to explore quantum-enhanced methods for experimental HEP data analysis.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Scaling the QUBO formulation to higher-dimensional or real experimental data could reveal advantages in handling complex systematic uncertainties.
Hybrid solvers might offer speedups for large-scale unfolding problems where matrix methods become computationally intensive.
Direct comparison on actual collider datasets would test whether the quantum-compatible encoding preserves fidelity under realistic noise.

Load-bearing premise

The QUBO encoding and hybrid solver represent the original unfolding objective faithfully without introducing artifacts that degrade performance beyond the synthetic tests shown.

What would settle it

Running the method on a realistic detector response matrix from actual high-energy physics data and finding statistically significant deviations from the known true distribution in regions not covered by the synthetic benchmarks.

Figures

Figures reproduced from arXiv: 2602.22776 by Carla Rieger, Gianluca Bianco, Marco Lorusso, Michele Grossi, Simone Gasperini.

read the original abstract

In experimental High-Energy Physics, unfolding refers to the problem of estimating the underlying distribution of a physical observable from detector-level data, in the presence of statistical fluctuations and systematic uncertainties. Starting from its reformulation as a regularized quadratic optimization problem, we develop a framework to address unfolding using both classical and quantum-compatible methods. In particular, we derive a Quadratic Unconstrained Binary Optimization (QUBO) representation of the unfolding objective, allowing direct implementation on quantum annealing and hybrid quantum-classical solvers. The proposed approach is implemented in QUnfold, an open-source Python package integrating classical mixed-integer solvers and D-Wave's hybrid quantum solver. We benchmark the method against widely used unfolding techniques in RooUnfold, including response Matrix Inversion, Iterative Bayesian Unfolding, and Singular Value Decomposition unfolding, using synthetic dataset with controlled distortion effects. Our results demonstrate that the optimization-based approach achieves competitive reconstruction accuracy across multiple distributions while naturally accommodating regularization within the objective function. This work establishes a unified optimization perspective on unfolding and provides a practical pathway for exploring quantum-enhanced methods in experimental HEP data analysis.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

They recast unfolding as a QUBO so it runs on quantum annealers and ship an open package, but the evidence stays on synthetic data with no direct check for encoding artifacts.

read the letter

The core move is turning the regularized quadratic unfolding objective into a QUBO that quantum-hybrid solvers can handle directly. That is new relative to the RooUnfold toolkit they cite, and they back it with an open Python package that also supports classical mixed-integer solvers. On the synthetic tests the method lands competitive accuracy with the usual matrix inversion, Bayesian, and SVD approaches while folding regularization into the objective without extra steps. That is useful for anyone who already thinks in optimization terms and wants a single framework that can later plug into quantum hardware. The soft spots are straightforward. All benchmarks use controlled synthetic distortions; there is no real detector data, no explicit propagation of systematic uncertainties, and no side-by-side run against a continuous quadratic solver to confirm the binary encoding plus penalty terms do not shift the recovered minimum. The stress-test worry about discretization artifacts on fine bins or low-count regions therefore remains open. The paper is aimed at the intersection of experimental HEP analysis and early quantum methods. A reader already working on unfolding or scouting quantum applications will get a concrete starting point and code to try. It is coherent on its own terms and shows clear thinking about the mapping, so it deserves a serious referee rather than a desk reject. I would send it out for review with the expectation that the authors add real-data tests and a continuous baseline comparison.

Referee Report

2 major / 2 minor

Summary. The manuscript reformulates unfolding in high-energy physics as a regularized quadratic minimization problem over bin contents, derives a QUBO encoding suitable for quantum annealing and hybrid solvers, implements the approach in the open-source QUnfold package, and benchmarks it against RooUnfold methods (matrix inversion, iterative Bayesian, SVD) on synthetic datasets with controlled distortions, claiming competitive reconstruction accuracy while naturally incorporating regularization.

Significance. If the central results hold, the work supplies a unified optimization perspective on unfolding that embeds regularization directly in the objective and supplies a practical route to quantum-compatible solvers via an open-source tool. The synthetic benchmarks and reproducible implementation are clear strengths that could facilitate follow-on studies in quantum-enhanced HEP analysis.

major comments (2)

[Methods (QUBO derivation)] Methods section on QUBO encoding: the binary-variable representation of continuous bin contents together with quadratic penalty terms for non-negativity and normalization is introduced without a direct numerical comparison to the solution of the original continuous quadratic program on the same response matrices; this leaves open the possibility that discretization or imperfect penalty scaling shifts the recovered minimum, especially for fine binning or low-count bins.
[Results (benchmarks)] Results section (synthetic benchmarks): the reported competitive accuracy is shown only on controlled synthetic distributions; the absence of quantitative error bars, real experimental data, and explicit propagation of systematic uncertainties means the claim that the method is ready for practical HEP use rests on an assumption that has not yet been tested.

minor comments (2)

[Abstract] Abstract: the statement of 'competitive reconstruction accuracy' would be strengthened by citing at least one concrete metric (e.g., average χ² or bias) from the benchmark tables.
[Methods] Notation: the transition from the continuous quadratic objective to the QUBO Hamiltonian should include an explicit equation showing how the regularization strength λ enters the final QUBO coefficients.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed review. We address each major comment below, indicating planned revisions where appropriate.

read point-by-point responses

Referee: Methods section on QUBO encoding: the binary-variable representation of continuous bin contents together with quadratic penalty terms for non-negativity and normalization is introduced without a direct numerical comparison to the solution of the original continuous quadratic program on the same response matrices; this leaves open the possibility that discretization or imperfect penalty scaling shifts the recovered minimum, especially for fine binning or low-count bins.

Authors: We agree that a direct numerical comparison would provide stronger validation of the QUBO approximation. In the revised manuscript we will add a dedicated comparison (new figure and text in the Methods section) between the continuous quadratic program solved via standard solvers and the QUBO formulation on the identical response matrices used in the benchmarks. This will quantify any shifts in the recovered minimum for both fine and coarse binning as well as low-count regimes, allowing readers to assess the impact of discretization and penalty scaling. revision: yes
Referee: Results section (synthetic benchmarks): the reported competitive accuracy is shown only on controlled synthetic distributions; the absence of quantitative error bars, real experimental data, and explicit propagation of systematic uncertainties means the claim that the method is ready for practical HEP use rests on an assumption that has not yet been tested.

Authors: The manuscript deliberately restricts the benchmarks to synthetic data with known ground truth to enable controlled, reproducible evaluation of the optimization framework. We do not claim the method is ready for practical HEP use; the abstract and conclusions describe competitive performance on synthetics and a practical pathway for quantum-compatible solvers. In revision we will (i) add quantitative error bars to all accuracy metrics, (ii) explicitly state the synthetic scope and limitations in the Results and Conclusions sections, and (iii) note that real-data validation and systematic-uncertainty propagation remain important future work. These clarifications will prevent any overstatement of readiness. revision: partial

Circularity Check

0 steps flagged

No circularity: standard reformulation of unfolding objective into QUBO

full rationale

The paper begins from the established regularized quadratic minimization formulation of unfolding (a standard approach in HEP literature), then applies the known binary encoding technique to obtain a QUBO representation. No equation reduces the final result to a fitted parameter or input by construction, no load-bearing self-citation is used to justify uniqueness or the central mapping, and the benchmarks against RooUnfold methods provide external comparison. The derivation chain is self-contained and does not rely on renaming or smuggling ansatzes from prior author work.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The framework rests on the assumption that unfolding can be expressed as a regularized quadratic program whose QUBO encoding preserves the original objective; no new physical entities are introduced.

free parameters (1)

regularization strength
Tuned parameter in the quadratic objective that controls smoothness of the unfolded distribution.

axioms (1)

domain assumption The detector response can be accurately captured by a response matrix relating true and observed distributions
Standard premise of all unfolding methods invoked at the start of the reformulation.

pith-pipeline@v0.9.0 · 5496 in / 1138 out tokens · 51384 ms · 2026-05-15T19:22:11.689535+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

27 extracted references · 27 canonical work pages · 1 internal anchor

[1]

Cambridge University Press, 2000

Rudolf Frühwirth and Meinhard Regler.Data analysis techniques for high-energy physics, volume 11. Cambridge University Press, 2000

work page 2000
[2]

An Unfolding Method for High Energy Physics Experiments

V olker Blobel. An unfolding method for high energy physics experiments.arXiv preprint hep-ex/0208022, 2002

work page internal anchor Pith review Pith/arXiv arXiv 2002
[3]

Rene Brun and Fons Rademakers. Root—an object oriented data analysis framework.Nuclear instruments and methods in physics research section A: accelerators, spectrometers, detectors and associated equipment, 389(1-2):81–86, 1997

work page 1997
[4]

John Wiley & Sons, 2010

Glenn F Knoll.Radiation detection and measurement. John Wiley & Sons, 2010

work page 2010
[5]

Data unfolding methods in high energy physics

Stefan Schmitt. Data unfolding methods in high energy physics. InEPJ web of conferences, volume 137, page 11008. EDP Sciences, 2017

work page 2017
[6]

Unfolding as Quantum Annealing.Journal of High Energy Physics, 2019(11):128, November 2019

Kyle Cormier, Riccardo Di Sipio, and Peter Wittek. Unfolding as Quantum Annealing.Journal of High Energy Physics, 2019(11):128, November 2019. arXiv:1908.08519 [hep-ex, physics:physics, physics:quant-ph]

work page arXiv 2019
[7]

Quantum entanglement and bell inequality violation in semi-leptonic top decays.Journal of High Energy Physics, 2024(7):1–43, 2024

Tao Han, Matthew Low, and Tong Arthur Wu. Quantum entanglement and bell inequality violation in semi-leptonic top decays.Journal of High Energy Physics, 2024(7):1–43, 2024

work page 2024
[8]

Comparison of unfolding methods using roofitunfold, 2020

Lydia Brenner, Pim Verschuuren, Rahul Balasubramanian, Carsten Burgard, Vincent Croft, Glen Cowan, and Wouter Verkerke. Comparison of unfolding methods using roofitunfold, 2020

work page 2020
[9]

Comparison of unfolding methods using roofitunfold.International Journal of Modern Physics A, 35(24):2050145, 2020

Lydia Brenner, Rahul Balasubramanian, Carsten Burgard, Wouter Verkerke, Glen Cowan, Pim Verschuuren, and Vincent Croft. Comparison of unfolding methods using roofitunfold.International Journal of Modern Physics A, 35(24):2050145, 2020

work page 2020
[10]

Omnifold: a method to simultaneously unfold all observables.Physical review letters, 124(18):182001, 2020

Anders Andreassen, Patrick T Komiske, Eric M Metodiev, Benjamin Nachman, and Jesse Thaler. Omnifold: a method to simultaneously unfold all observables.Physical review letters, 124(18):182001, 2020

work page 2020
[11]

The landscape of unfolding with machine learning.arXiv preprint arXiv:2404.18807, 2024

Nathan Huetsch, Javier Mariño Villadamigo, Alexander Shmakov, Sascha Diefenbacher, Vinicius Mikuni, Theo Heimel, Michael Fenton, Kevin Greif, Benjamin Nachman, Daniel Whiteson, et al. The landscape of unfolding with machine learning.arXiv preprint arXiv:2404.18807, 2024. 7

work page arXiv 2024
[12]

Event-by-event comparison between machine-learning-and transfer-matrix-based unfolding methods.The European Physical Journal C, 84(8):770, 2024

Mathias Backes, Anja Butter, Monica Dunford, and Bogdan Malaescu. Event-by-event comparison between machine-learning-and transfer-matrix-based unfolding methods.The European Physical Journal C, 84(8):770, 2024

work page 2024
[13]

Quantum computing for high-energy physics: State of the art and challenges.PRX Quantum, 5(3):037001, 2024

Alberto Di Meglio, Karl Jansen, Ivano Tavernelli, Constantia Alexandrou, Srinivasan Arunachalam, Christian W Bauer, Kerstin Borras, Stefano Carrazza, Arianna Crippa, Vincent Croft, et al. Quantum computing for high-energy physics: State of the art and challenges.PRX Quantum, 5(3):037001, 2024

work page 2024
[14]

Study and application of qunfold, an innovative quantum annealing-based unfolding tool, using simulated atlas data

Valerio Brugnami, Simone Gasperini, Marco Lorusso, Gianluca Bianco, and Matteo Franchini. Study and application of qunfold, an innovative quantum annealing-based unfolding tool, using simulated atlas data. In2025 International Conference on Quantum Communications, Networking, and Computing (QCNC), pages 606–613, 2025

work page 2025
[15]

Oxford university press, 1998

Glen Cowan.Statistical data analysis. Oxford university press, 1998

work page 1998
[16]

Svd approach to data unfolding.Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 372(3):469–481, 1996

Andreas Hoecker and Vakhtang Kartvelishvili. Svd approach to data unfolding.Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 372(3):469–481, 1996

work page 1996
[17]

Giulio D’Agostini. A multidimensional unfolding method based on bayes’ theorem.Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 362(2-3):487–498, 1995

work page 1995
[18]

Solution of incorrectly formulated problems and the regularization method.Sov Dok, 4:1035–1038, 1963

Andrei N Tikhonov. Solution of incorrectly formulated problems and the regularization method.Sov Dok, 4:1035–1038, 1963

work page 1963
[19]

Maximum likelihood from incomplete data via the em algorithm.Journal of the royal statistical society: series B (methodological), 39(1):1–22, 1977

Arthur P Dempster, Nan M Laird, and Donald B Rubin. Maximum likelihood from incomplete data via the em algorithm.Journal of the royal statistical society: series B (methodological), 39(1):1–22, 1977

work page 1977
[20]

Quadratic knapsack problems.Combinatorial Optimization, pages 132–149, 1980

G Gallo, PL Hammer, and B Simeone. Quadratic knapsack problems.Combinatorial Optimization, pages 132–149, 1980

work page 1980
[21]

A nonlinear programming algorithm for solving semidefinite programs via low-rank factorization.Mathematical programming, 95(2):329–357, 2003

Samuel Burer and Renato DC Monteiro. A nonlinear programming algorithm for solving semidefinite programs via low-rank factorization.Mathematical programming, 95(2):329–357, 2003

work page 2003
[22]

Integer quadratic optimization.European journal of operational research, 19(2):268–273, 1985

Frank Körner. Integer quadratic optimization.European journal of operational research, 19(2):268–273, 1985

work page 1985
[23]

D-wave hybrid solver service: An overview

D-Wave Systems Inc. D-wave hybrid solver service: An overview. Technical Report 14-1039A-B, D-Wave Systems Inc., May 2020

work page 2020
[24]

The unconstrained binary quadratic programming problem: a survey.Journal of combinatorial optimization, 28:58–81, 2014

Gary Kochenberger, Jin-Kao Hao, Fred Glover, Mark Lewis, Zhipeng Lü, Haibo Wang, and Yang Wang. The unconstrained binary quadratic programming problem: a survey.Journal of combinatorial optimization, 28:58–81, 2014

work page 2014
[25]

Ising formulations of many np problems.Frontiers in physics, 2:5, 2014

Andrew Lucas. Ising formulations of many np problems.Frontiers in physics, 2:5, 2014

work page 2014
[26]

Quantum bridge analytics i: a tutorial on formulating and using qubo models.4or, 17(4):335–371, 2019

Fred Glover, Gary Kochenberger, and Yu Du. Quantum bridge analytics i: a tutorial on formulating and using qubo models.4or, 17(4):335–371, 2019

work page 2019
[27]

Mniszewski, and Christian F

Benjamin Krakoff, Susan M. Mniszewski, and Christian F. A. Negre. Controlled precision QUBO-based algorithm to compute eigenvectors of symmetric matrices.PLOS ONE, 17(5):e0267954, May 2022. 8

work page 2022