pith. sign in

arxiv: 2402.09848 · v2 · pith:OQBOA22Cnew · submitted 2024-02-15 · 🪐 quant-ph

Parameterized quantum circuits as universal generative models for continuous multivariate distributions

Alice Barthe , Michele Grossi , Sofia Vallecorsa , Jordi Tura , Vedran Dunjko This is my paper

Pith reviewed 2026-05-24 03:42 UTC · model grok-4.3

classification 🪐 quant-ph
keywords parameterized quantum circuitsgenerative modelsuniversalitycontinuous distributionsmultivariate distributionsexpectation value samplingvariational quantum algorithmsquantum machine learning
0
0 comments X

The pith

Parameterized quantum circuits can universally generate any continuous multivariate distribution by sampling expectation values after uploading classical random data.

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

The paper proves that expectation-value sampling models based on parameterized quantum circuits are universal generators for distributions over continuous multivariate variables. Classical random inputs are fed into the circuit, and the outputs are expectation values of a fixed set of observables. A reader would care because this supplies the missing universality guarantee for quantum generative modeling on continuous data, where prior results were limited to supervised tasks. The authors also derive tight bounds linking the smallest number of qubits and measurements that still permit universality across different circuit architectures.

Core claim

Expectation value sampling-based parameterized quantum circuits, which upload classical random data and return the expectation values of a fixed set of observables, are universal for the generation of any multivariate distribution over continuous variables. Multiple circuit architectures achieve this universality, and the paper supplies tight bounds that relate the minimal required qubit number to the minimal required number of measurements.

What carries the argument

Expectation value sampling from parameterized quantum circuits after classical random data upload, with fixed observables whose expectations form the output distribution.

If this is right

  • Specific circuit architectures suffice to reach universality for arbitrary continuous multivariate distributions.
  • The minimal number of qubits needed is bounded from below in a way that depends on the target distribution dimension.
  • The minimal number of distinct observables (hence measurements) needed is also tightly bounded.
  • These bounds can directly constrain the resource requirements when designing quantum circuits for generative tasks.

Where Pith is reading between the lines

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

  • The same upload-and-expectation scheme might be tested on discrete distributions to see whether the universality proof adapts.
  • Practical noise in real quantum hardware could be checked against the derived qubit and measurement lower bounds to estimate when the universality becomes unreachable.
  • The fixed-observable restriction suggests that learning the observable set itself, rather than only the circuit parameters, might enlarge the model class.

Load-bearing premise

Universality is claimed only for models that output expectation values of a fixed set of observables after uploading classical random data into the circuit.

What would settle it

Existence of even one continuous multivariate distribution that cannot be approximated arbitrarily well by the expectation values of any fixed set of observables from any parameterized quantum circuit that receives classical random inputs.

Figures

Figures reproduced from arXiv: 2402.09848 by Alice Barthe, Jordi Tura, Michele Grossi, Sofia Vallecorsa, Vedran Dunjko.

Figure 1
Figure 1. Figure 1: FIG. 1: Expectation value sampling model: a random vec [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3: Dense Encoding Circuit as a universal generator, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4: Visual summary of results. We show the asymptotic [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5: A sequence of continuous functions converging point [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
read the original abstract

Parameterized quantum circuits have been extensively used as the basis for machine learning models in regression, classification, and generative tasks. For supervised learning, their expressivity has been thoroughly investigated and several universality properties have been proven. However, in the case of quantum generative modelling, much less is known, especially when the task is to model distributions over continuous variables. In this work, we elucidate expectation value sampling-based models. Such models output the expectation values of a set of fixed observables from a quantum circuit into which classical random data has been uploaded. We prove the universality of such variational quantum algorithms for the generation of multivariate distributions. We explore various architectures which allow universality and prove tight bounds connecting the minimal required qubit number, and the minimal required number of measurements needed. Our results may help guide the design of future quantum circuits in generative modelling tasks.

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.

Referee Report

0 major / 1 minor

Summary. The manuscript proves that parameterized quantum circuits operating under an expectation-value sampling scheme—classical random inputs encoded into the circuit with outputs given by expectations of a fixed finite set of observables—are universal approximators for continuous multivariate probability distributions. It supplies explicit circuit architectures that achieve this universality and derives matching upper and lower bounds on the minimal qubit count and measurement count required.

Significance. If the stated universality result and bounds hold, the work supplies a concrete theoretical foundation for quantum generative models of continuous variables, extending prior universality results from supervised settings. The explicit constructions together with tight resource bounds constitute a clear strength that can directly inform circuit design; the reliance on standard density arguments over compact sets keeps the derivations within conventional analytic tools.

minor comments (1)
  1. [Abstract] Abstract: the scope of the universality claim (restricted to expectation-value sampling with a fixed observable set) could be stated more explicitly to prevent readers from extrapolating beyond the model class analyzed.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their careful reading and positive assessment of the manuscript. We are pleased that the universality result, explicit constructions, and tight resource bounds were found to be clear strengths.

Circularity Check

0 steps flagged

Derivation is self-contained; no circular steps identified

full rationale

The manuscript proves universality for a restricted class of expectation-value sampling models by supplying explicit PQC architectures and deriving tight qubit/measurement bounds via standard density arguments on compact sets. No load-bearing step reduces to a fitted input renamed as prediction, a self-citation chain, or an ansatz smuggled from prior work by the same authors. The argument is scoped explicitly to the model class and remains internally consistent without invoking unverified uniqueness theorems or self-definitional equivalences.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the definition of expectation-value sampling models and standard quantum circuit assumptions; no free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption Expectation values of a fixed set of observables after uploading classical random data can serve as the output of a generative model.
    This defines the precise class of models for which universality is claimed.

pith-pipeline@v0.9.0 · 5678 in / 1048 out tokens · 38084 ms · 2026-05-24T03:42:02.208051+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Quantum generative modeling for financial time series with temporal correlations

    quant-ph 2025-07 unverdicted novelty 5.0

    QGANs with quantum generators and classical discriminators generate financial time series matching target distributions and desired temporal correlations, with quality varying by circuit depth, bond dimension, and sim...

Reference graph

Works this paper leans on

26 extracted references · 26 canonical work pages · cited by 1 Pith paper · 1 internal anchor

  1. [1]

    merlin.mbs apsrev4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked

    FUNCTION id.bst "merlin.mbs apsrev4-1.bst 2010-07-25 4.21a (PWD, AO, DPC) hacked" ENTRY address archive archivePrefix author bookaddress booktitle chapter collaboration doi edition editor eid eprint howpublished institution isbn issn journal key language month note number organization pages primaryClass publisher school SLACcitation series title translati...

    work page 2010

  2. [2]

    The learnability of quantum states

    Scott Aaronson. The learnability of quantum states. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences , 463(2088):3089--3114, September 2007. Publisher: Royal Society

    work page 2088

  3. [3]

    Amin, Evgeny Andriyash, Jason Rolfe, Bohdan Kulchytskyy, and Roger Melko

    Mohammad H. Amin, Evgeny Andriyash, Jason Rolfe, Bohdan Kulchytskyy, and Roger Melko. Quantum Boltzmann Machine . Physical Review X , 8(2):021050, May 2018. Publisher: American Physical Society

    work page 2018

  4. [4]

    Gradient Flows

    Luigi Ambrosio, Nicola Gigli, and Giuseppe Savaré. Gradient Flows . Lectures in Mathematics ETH Zürich . Birkhäuser-Verlag, Basel, 2005

    work page 2005

  5. [5]

    Dense quantum coding and quantum finite automata

    Andris Ambainis, Ashwin Nayak, Amnon Ta-Shma, and Umesh Vazirani. Dense quantum coding and quantum finite automata. Journal of the ACM , 49(4):496--511, July 2002

    work page 2002

  6. [6]

    R. Baire. Sur les fonctions de variables réelles. Annali di Matematica Pura ed Applicata (1898-1922) , 3(1):1--123, December 1899

    work page 1922

  7. [7]

    Kottmann, Tim Menke, Wai-Keong Mok, Sukin Sim, Leong-Chuan Kwek, and Alán Aspuru-Guzik

    Kishor Bharti, Alba Cervera-Lierta, Thi Ha Kyaw, Tobias Haug, Sumner Alperin-Lea, Abhinav Anand, Matthias Degroote, Hermanni Heimonen, Jakob S. Kottmann, Tim Menke, Wai-Keong Mok, Sukin Sim, Leong-Chuan Kwek, and Alán Aspuru-Guzik. Noisy intermediate-scale quantum algorithms. Reviews of Modern Physics , 94(1):015004, February 2022. Publisher: American Phy...

    work page 2022

  8. [8]

    V. I. Bogachev, A. V. Kolesnikov, and K. V. Medvedev. Triangular transformations of measures. Sbornik: Mathematics , 196(3):309, April 2005. Publisher: IOP Publishing

    work page 2005

  9. [9]

    Cerezo, Andrew Arrasmith, Ryan Babbush, Simon C

    M. Cerezo, Andrew Arrasmith, Ryan Babbush, Simon C. Benjamin, Suguru Endo, Keisuke Fujii, Jarrod R. McClean, Kosuke Mitarai, Xiao Yuan, Lukasz Cincio, and Patrick J. Coles. Variational quantum algorithms. Nature Reviews Physics , 3(9):625--644, September 2021. Number: 9 Publisher: Nature Publishing Group

    work page 2021

  10. [10]

    in preparation

    Kevin Chen, Andrii Kurkin, Hao Wang, and Vedran Dunjko. in preparation

    work page

  11. [11]

    The Born supremacy: quantum advantage and training of an Ising Born machine

    Brian Coyle, Daniel Mills, Vincent Danos, and Elham Kashefi. The Born supremacy: quantum advantage and training of an Ising Born machine. npj Quantum Information , 6(1):1--11, July 2020. Number: 1 Publisher: Nature Publishing Group

    work page 2020

  12. [12]

    Quantum generative adversarial networks

    Pierre-Luc Dallaire-Demers and Nathan Killoran. Quantum generative adversarial networks. Physical Review A , 98(1):012324, July 2018. Publisher: American Physical Society

    work page 2018

  13. [13]

    Neural Autoregressive Flows

    Chin-Wei Huang, David Krueger, Alexandre Lacoste, and Aaron Courville. Neural Autoregressive Flows . In Proceedings of the 35th International Conference on Machine Learning , pages 2078--2087. PMLR, July 2018. ISSN: 2640-3498

    work page 2078

  14. [14]

    Marshall, Riccardo Molteni, and Vedran Dunjko

    Sofiene Jerbi, Casper Gyurik, Simon C. Marshall, Riccardo Molteni, and Vedran Dunjko. Shadows of quantum machine learning, May 2023. arXiv:2306.00061 [quant-ph, stat]

    work page arXiv 2023

  15. [15]

    Sum-of-Squares Polynomial Flow

    Priyank Jaini, Kira A. Selby, and Yaoliang Yu. Sum of Squares Polynomial Flow , June 2019. arXiv:1905.02325 [cs, stat]

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  16. [16]

    Kobyzev, S

    I. Kobyzev, S. D. Prince, and M. A. Brubaker. Normalizing Flows : An Introduction and Review of Current Methods . IEEE Transactions on Pattern Analysis & Machine Intelligence , 43(11):3964--3979, November 2021. Place: Los Alamitos, CA, USA Publisher: IEEE Computer Society

    work page 2021

  17. [17]

    Laczkovich

    M. Laczkovich. Baire 1 functions. Real Analysis Exchange , 9(1):15--28, January 1984. Publisher: Michigan State University Press

    work page 1984

  18. [18]

    Lebesgue

    H. Lebesgue. Une propriété caractéristique des fonctions de classe 1. Bulletin de la Société Mathématique de France , 32:229--242, 1904. Publisher: Société mathématique de France

    work page 1904

  19. [19]

    Differentiable learning of quantum circuit Born machines

    Jin-Guo Liu and Lei Wang. Differentiable learning of quantum circuit Born machines. Physical Review A , 98(6):062324, December 2018. Publisher: American Physical Society

    work page 2018

  20. [20]

    A Variational Algorithm for Quantum Neural Networks

    Antonio Macaluso, Luca Clissa, Stefano Lodi, and Claudio Sartori. A Variational Algorithm for Quantum Neural Networks . In Valeria V. Krzhizhanovskaya, Gábor Závodszky, Michael H. Lees, Jack J. Dongarra, Peter M. A. Sloot, Sérgio Brissos, and João Teixeira, editors, Computational Science – ICCS 2020 , Lecture Notes in Computer Science , pages 591--604, Ch...

    work page 2020

  21. [21]

    Quantum-state preparation with universal gate decompositions

    Martin Plesch and Časlav Brukner. Quantum-state preparation with universal gate decompositions. Physical Review A , 83(3):032302, March 2011. Publisher: American Physical Society

    work page 2011

  22. [22]

    Forn-Díaz, and José I

    Adrián Pérez-Salinas, David López-Núñez, Artur García-Sáez, P. Forn-Díaz, and José I. Latorre. One qubit as a Universal Approximant . Physical Review A , 104(1):012405, July 2021. arXiv:2102.04032 [quant-ph]

    work page arXiv 2021

  23. [23]

    Variational Quantum Generators : Generative Adversarial Quantum Machine Learning for Continuous Distributions

    Jonathan Romero and Alán Aspuru-Guzik. Variational Quantum Generators : Generative Adversarial Quantum Machine Learning for Continuous Distributions . Advanced Quantum Technologies , 4(1):2000003, 2021

    work page 2021

  24. [24]

    Effect of data encoding on the expressive power of variational quantum-machine-learning models

    Maria Schuld, Ryan Sweke, and Johannes Jakob Meyer. Effect of data encoding on the expressive power of variational quantum-machine-learning models. Physical Review A , 103(3):032430, March 2021. Publisher: American Physical Society

    work page 2021

  25. [25]

    Optimal Transport , volume 338 of Grundlehren der mathematischen Wissenschaften

    Cédric Villani. Optimal Transport , volume 338 of Grundlehren der mathematischen Wissenschaften . Springer, Berlin, Heidelberg, 2009

    work page 2009

  26. [26]

    Numerical Methods for Stochastic Computations : A Spectral Method Approach

    Dongbin Xiu. Numerical Methods for Stochastic Computations : A Spectral Method Approach . Princeton University Press, 2010

    work page 2010