Pith. sign in

REVIEW 3 major objections 6 minor 62 references

New Quantum Algorithm for Principal Component Analysis

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims a quantum PCA algorithm whose sample and circuit cost depend on the gap between the two largest eigenvalues rather than on the largest eigenvalue itself.

desk verdict A gap-based QPCA variant with a nice regime insight, but the keystone block-encoding step conflates a CPTP channel with a unitary and the covariance preparation mishandles its garbage term. read the letter →

arxiv 2501.07891 v1 pith:KHCI6NPS submitted 2025-01-14 quant-ph

Nhat A. Nghiem This is my paper · ORCID
classification quant-ph MSC 81P6868Q12 PACS 03.67.Ac03.67.Lx
keywords quantumprincipalcomponentanalysissingularvaluetransformationblockencodingdensitymatrixexponentiationpowermethodcovariancepreparationeigenvalueestimation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper is trying to establish a new quantum algorithm for principal component analysis that works directly from copies of a quantum state. It claims that, for a state with largest eigenvalue $r_1$ and gap $\gamma$ below the second-largest eigenvalue, the largest eigenpair can be recovered to error $\epsilon$ using only $O(\gamma^{-2}\log^2(1/\epsilon)\epsilon^{-2})$ copies of the state and a circuit of depth $O(\log(n)\gamma^{-3}\log^3(1/\epsilon)\epsilon^{-2})$. This makes the algorithm's cost set by the gap rather than by the size of the dominant eigenvalue, which is exactly the regime where the original quantum PCA is slow. The paper also gives a block-encoding protocol for preparing a centered covariance matrix from classical data, including the centroid term, which earlier covariance-preparation work left out.

What carries the argument

The carrying object is the block encoding: a larger unitary whose top-left block equals (or approximates) the matrix one wants to manipulate. Lemma 4 is the load-bearing construction: it takes copies of $\rho$, uses density-matrix exponentiation to approximate $\exp(-i\rho/2)$, and then uses the quantum-singular-value-transformation logarithm-of-unitary lemma to convert that unitary into an $\epsilon$-approximated block encoding of $\pi\rho/4$. Once the block encoding exists, Lemma 6's quantum power method extracts the largest eigenpair, and the linear-combination and scaling lemmas assemble block encodings of covariance matrices and of $\rho-r_1|\lambda_1\rangle\langle\lambda_1|$ for iterative extraction of further principal components.

What would settle it

Perform quantum process tomography on the circuit that Lemma 4 claims to be an $\epsilon$-approximated block encoding of $\pi\rho/4$ for a small known $\rho$; if the induced map on the signal register differs from $\pi\rho/4$ by more than $\epsilon$ in the appropriate norm, Theorem 1's foundation fails. A complementary check is to compute the Choi matrix of the channel in Lemma 1: a unitary $U$ exists only if that channel is rank one.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1: given multiple copies of a density matrix $\rho\in\mathbb{C}^{n\times n}$ whose two largest eigenvalues are separated by $\gamma=|r_1-r_2|$, there is a quantum algorithm that outputs an estimate $\tilde r$ and a state $|\tilde x\rangle$ with $|\tilde r-r_1|\le\epsilon$ and $\bigl||\tilde x\rangle-|\lambda_1\rangle\bigr|_2\le\epsilon$, using $N=O(\gamma^{-2}\log^2(1/\epsilon)\epsilon^{-2})$ copies of $\rho$ and circuit depth $O(\log(n)\gamma^{-3}\log^3(1/\epsilon)\epsilon^{-2})$. The route is to turn $\rho$ into a block encoding of $\pi\rho/4$, feed that block encoding into a quantum power method, and iteratively subtract discovered components to get the next principal components. The paper further claims two procedures that prepare the centered covariance matrix $\frac{\pi}{8}(\bar\rho-\mu\mu^\dagger)$ from classical data, one approximate and one exact, thereby incorporating the centroid that is ignored by 'PCA without centering'.

Load-bearing premise

The load-bearing premise is that the density-matrix exponentiation procedure of Lemma 1 gives a unitary that can play the role of $U$ in the QSVT logarithm-of-unitary lemma, but as stated that procedure is a channel obtained by tracing out the copies of $\rho$, and the paper does not show how to make it coherent; if that step fails, the block encoding of $\pi\rho/4$ and the whole algorithm do not exist.

Editorial extensions

If this is right

  • If Theorem 1 holds, the dominant-eigenvalue problem for $\rho$ is solvable with $1/\epsilon^2$ scaling in error and $1/\gamma^2$ scaling in gap, while the original QPCA has $1/(r_1^2\epsilon^3)$ scaling; the new algorithm wins when $\gamma$ is large even if $r_1$ is small.
  • The two QPCA frameworks complement each other: original QPCA is best when the top eigenvalues are nearly equal and dominate the spectrum, while the new one is best when the gap between the top two eigenvalues is $O(1)$, including high-rank states.
  • Theorem 2 extends the algorithm to the top $R$ principal components, but the cost grows exponentially in $R$, so the advantage is for small $R$.
  • The covariance-matrix preparation lemma gives an exact block encoding of $\frac{\pi}{8}(\bar\rho-\mu\mu^\dagger)$ at depth $O(N+N\log n)$, so for small datasets the covariance preparation step adds no approximation error.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the block-encoding step can be made coherent, the same gap-dependent machinery would apply to other spectral tasks, such as estimating the gap itself or testing whether a state is low rank; the paper does not explore these.
  • The logarithmic dependence on dimension and linear dependence on number of data points suggest the protocol is aimed at high-dimensional, small-sample datasets, a domain the paper mentions but does not quantify with explicit bounds.
  • A direct numerical test on synthetic covariance matrices with controlled gap $\gamma$ could reveal the constant factors hidden by the big-$O$ notation.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The manuscript proposes a new quantum algorithm for principal component analysis (QPCA). The central idea is to use density matrix exponentiation to obtain an approximation to exp(-iρt), then use the quantum singular value transformation (QSVT) framework to construct a block encoding of πρ/4, and finally apply a quantum power method (Lemma 6) to extract the largest eigenvalue and eigenvector of ρ. The claimed complexity of Theorem 1 depends on the spectral gap γ between the two largest eigenvalues and improves on the original QPCA in certain regimes. The paper also proposes a method for preparing a covariance matrix from classical data, including the contribution from the centroid μμ†, using block-encoding techniques. The manuscript is primarily a complexity-theoretic argument with no numerical simulations or machine-checked proofs.

Significance. If the central claims were valid, the algorithm would be a useful complement to the original QPCA, particularly when the largest eigenvalue gap is large, and the covariance-matrix preparation would address the centroid term that was not fully handled in prior work. However, the significance is undermined by a fundamental technical gap: the procedure cited as implementing exp(-iρt) actually implements a non-unitary quantum channel, and the paper does not provide a coherent unitary block encoding of πρ/4. Additionally, the covariance-matrix construction contains a concrete normalization error. The paper honestly compares its complexity with existing work, but the main theorem and the covariance-preparation lemma are not established by the presented arguments.

major comments (3)
  1. [Section III, Lemma 4 and Theorem 1] The construction of the block encoding of πρ/4 relies on Lemma 1 providing a unitary U = exp(-iρ/2) that can be used in Lemma 2 (logarithm of a unitary). However, Lemma 1, as derived in Section II (Eqs. (1)-(3)), implements the CPTP channel Λ_t(σ) = Tr_1[e^{-iS t}(ρ⊗σ)e^{iS t}], which is not a unitary conjugation on the target system whenever ρ is mixed. For example, with ρ = I/2, the channel maps a pure state to a mixed state for t ≠ 0, whereas a unitary channel preserves purity. Lemma 2 explicitly requires a unitary U with controlled-U and U† as primitives. The paper does not provide a coefficient or an uncomputation argument to turn this channel into a coherent unitary acting on the target system alone. Consequently, the assertion in Section III that 'any unitary trivially block encodes itself' does not apply to the output of Lemma 1, and the block encoding of πρ/4 is not established. Since Theorem 1 feeds this purported block encoding into the quantum power method (Lemma 6), the central claim is unsupported.
  2. [Section V, Eqs. (14)-(17) and Lemma 8] The covariance-matrix preparation contains a concrete normalization error. In Eq. (14), Up is a unitary block encoding of Σ_i p_i U_i, so its action on |0⟩|0⟩ is |0⟩(Σ_i p_i |x_i⟩) + |Garbage⟩, where the total state is normalized. The vector |Φ⟩ defined in Eq. (15) as |0⟩(Σ_i p_i |x_i⟩) is generally not normalized, so no unitary can generate it. Moreover, Lemma 9, as stated in the manuscript, constructs a block encoding of the reduced density matrix ρ = Tr_A |Φ⟩⟨Φ|, not of the full projector |Φ⟩⟨Φ|; the text incorrectly says 'construct the block encoding of |Φ⟩⟨Φ|'. The reduced density matrix of the state produced by Up is not equal to μμ†, because the |Garbage⟩ component contributes additional terms. Therefore the claimed block encoding of μμ† and the resulting covariance-matrix preparation in Lemma 8 are not correct as written.
  3. [Section III, Lemma 6 and Theorem 1] The quantum power method Lemma 6, which is the core subroutine for extracting the largest eigenvalue and eigenvector, is quoted verbatim from the author's own preprints [45,50] and is not proved, derived, or independently verified in this manuscript. Theorem 1 inherits this dependence, so the paper's central claim is not self-contained. While it is common to build on prior work, the cited sources are not peer-reviewed, and the lemma statement is the main technical engine of the algorithm. The manuscript should either provide a full proof of Lemma 6 in an appendix or explicitly frame Theorem 1 as conditional on the validity of [45,50].
minor comments (6)
  1. [Section II, Eq. (1)] The second line of Eq. (1) writes 'exp(-ρΔt)' instead of 'exp(-iρΔt)'; this is a typo that could confuse the reader.
  2. [Section III, Lemma 3] The error term '4dp ε/α' in Lemma 3 appears dimensionally inconsistent; in the standard QSVT statement the error is proportional to d√(ε/α), and the bound |P(x)| ≤ 1/2 should be checked against the source. The manuscript later sets δ to a constant, but the dependence of the approximation error on δ is not made explicit.
  3. [Section V, Eq. (14)] The symbol M in '|0⟩^{⊗ log(M)}' is not defined; it should be the dimension of the ancilla or the number of qubits, and the notation is inconsistent with the rest of the paper.
  4. [Section V, Eq. (13)] The displayed form of Ui shows only the first column as |xi⟩; the remaining columns are not described, which makes the subsequent discussion of the 'first column of Σ p_i U_i' less clear.
  5. [General] The manuscript contains many typos and grammatical errors, such as 'alternatively new quantum framework' in the abstract and inconsistent use of † and T for adjoints. A thorough editing pass is needed.
  6. [References] References [45] and [50] are arXiv preprints. The manuscript should indicate whether these have been peer-reviewed, or include enough detail in an appendix so that the key lemma can be evaluated independently.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivation is a modular composition of external QSVT results and an independently-stated prior power-method lemma; the paper's serious gaps are correctness issues, not input-equivalent reductions.

full rationale

Theorem 1 is assembled from four modular ingredients: Lemma 1 (density matrix exponentiation, from Lloyd et al. [1]), Lemmas 2-3 (QSVT, from Gilyen et al. [43]), Lemma 4 (a composition of these), and Lemma 6 (the quantum power method). The only component cited from the author's own prior work is Lemma 6, which is stated as a parameter-free oracle-model theorem: given a block encoding of A with complexity T_A and a spectral gap γ, it outputs the largest eigenvector/eigenvalue with explicit complexity. Its assumptions do not include the target result of this paper (extracting principal components from copies of ρ), so under the review rules it counts as independent support rather than a self-citation loop. The covariance-matrix preparation in Section V is a constructive protocol, not a fitted parameter relabeled as a prediction, and no uniqueness theorem from the authors' prior work is used to force the construction. The most serious issues in the paper are correctness gaps, not circularity: Lemma 1's density-matrix-exponentiation procedure is a CPTP channel, not a unitary on the target system, so Lemma 2's requirement that U = exp(-iH) be a unitary is not met by the stated construction; and the factor in Lemma 4 ('πρ/4' versus Corollary 71's '2H/π') appears inconsistent. These could invalidate the derivation if unrepairable, but they are not cases where the conclusion is equivalent to the premise by construction. No circular step was found, so the circularity score is 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

No data fitting occurs, so there are no fitted free parameters. The listed axioms are the unproven premises the central results depend on: the unitary character of density matrix exponentiation, the black-box quantum power method from the author's prior work, and the erroneous Up-to-|Φ> step in the covariance matrix preparation.

assumptions (3)
  • ad hoc to paper Density matrix exponentiation can be implemented as a unitary approximation to exp(-iρt) suitable for QSVT.
    Lemma 1 describes a channel; the paper assumes a unitary without proof in Section III before Lemma 4.
  • domain assumption The quantum power method of Refs. [45,50] works with an ϵ-approximated block encoding and achieves the stated complexity.
    Lemma 6 is taken from the author's own prior work; no proof or independent check is given in this paper.
  • ad hoc to paper Up, a block encoding of Σ p_i U_i, generates the pure state |Φ> = |0> Σ p_i |x_i> from |0>|0>, enabling Lemma 9.
    Eq. (14)-(15) show a |Garbage> term is present, so Up does not prepare |Φ> exactly.

how reviews work

0 comments
Cite this review

Pith. "Pith review of New Quantum Algorithm for Principal Component Analysis." pith.science (2026). https://pith.science/paper/KHCI6NPS

@misc{pith2026250107891,
  author       = {Pith},
  title        = {Pith review of: New Quantum Algorithm for Principal Component Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KHCI6NPS}},
  note         = {Machine review of arXiv:2501.07891}
}
read the original abstract

Quantum principal component analysis (QPCA) ignited a new development toward quantum machine learning algorithms. Initially showcasing as an active way for analyzing a quantum system using the quantum state itself, QPCA also found potential application in analyzing a large-scale dataset, represented by the so-called covariance matrix. Inspired by recent advancement in quantum algorithms, we give an alternatively new quantum framework for performing principal component analysis. By analyzing the performance in detail, we shall identify the regime in which our proposal performs better than the original QPCA. In addition, we also provide a new approach for preparing the covariance matrix, given classical dataset, on a quantum computer. Thus, our work provides an efficient complementary framework for revealing features of the quantum state, while keeping the philosophy of original QPCA, as the state could play an active role in analyzing itself.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 25 canonical work pages

  1. [1]

    Quantum algorithms for supervised and unsupervised machine learning

    Seth Lloyd, Masoud Mohseni, and Patrick Rebentrost. Quantum algorithms for supervised and unsupervised machine learning. arXiv preprint arXiv:1307.0411, 2013

    arXiv 2013

  2. [2]

    Quantum measurements and the abelian stabilizer problem

    A Yu Kitaev. Quantum measurements and the abelian stabilizer problem. arXiv preprint quant-ph/9511026, 1995

    arXiv 1995

  3. [3]

    Architectures for a quantum random access memory

    Vittorio Giovannetti, Seth Lloyd, and Lorenzo Maccone. Architectures for a quantum random access memory. Physical Review A—Atomic, Molecular, and Optical Physics, 78(5):052310, 2008

    work page 2008

  4. [4]

    Quantum random access memory

    Vittorio Giovannetti, Seth Lloyd, and Lorenzo Maccone. Quantum random access memory. Physical review letters, 100(16):160501, 2008

    2008

  5. [5]

    Covariance matrix preparation for quantum principal component analysis

    Max Hunter Gordon, Marco Cerezo, Lukasz Cincio, and Patrick J Coles. Covariance matrix preparation for quantum principal component analysis. PRX Quantum, 3(3):030334, 2022

    work page 2022

  6. [6]

    A fast quantum mechanical algorithm for database search

    Lov K Grover. A fast quantum mechanical algorithm for database search. In Proceedings of the twenty-eighth annual ACM symposium on Theory of computing, pages 212–219, 1996

    1996

  7. [7]

    Polynomial-time algorithms for prime factorization and discrete logarithms on a quantum computer

    Peter W Shor. Polynomial-time algorithms for prime factorization and discrete logarithms on a quantum computer. SIAM review, 41(2):303–332, 1999

    1999

  8. [8]

    Quantum theory, the church–turing principle and the universal quantum computer

    David Deutsch. Quantum theory, the church–turing principle and the universal quantum computer. Proceedings of the Royal Society of London. A. Mathematical and Physical Sciences, 400(1818):97–117, 1985

    1985

Show all 62 references
  1. [9]

    Rapid solution of problems by quantum computation

    David Deutsch and Richard Jozsa. Rapid solution of problems by quantum computation. Proceedings of the Royal Society of London. Series A: Mathematical and Physical Sciences, 439(1907):553–558, 1992

    1907

  2. [10]

    Simulating physics with computers

    Richard P Feynman. Simulating physics with computers. In Feynman and computation, pages 133–153. CRC Press, 2018

    2018

  3. [11]

    Universal quantum simulators

    Seth Lloyd. Universal quantum simulators. Science, 273(5278):1073–1078, 1996

    1996

  4. [12]

    Efficient quantum algorithms for simulating sparse hamiltonians

    Dominic W Berry, Graeme Ahokas, Richard Cleve, and Barry C Sanders. Efficient quantum algorithms for simulating sparse hamiltonians. Communications in Mathematical Physics, 270(2):359–371, 2007

    2007

  5. [13]

    Black-box hamiltonian simulation and unitary implementation

    Dominic W Berry and Andrew M Childs. Black-box hamiltonian simulation and unitary implementation. Quantum Information and Computation, 12:29–62, 2009

    2009

  6. [14]

    High-order quantum algorithm for solving linear differential equations

    Dominic W Berry. High-order quantum algorithm for solving linear differential equations. Journal of Physics A: Mathe- matical and Theoretical, 47(10):105301, 2014

    2014

  7. [15]

    Hamiltonian simulation with nearly optimal dependence on all parameters

    Dominic W Berry, Andrew M Childs, and Robin Kothari. Hamiltonian simulation with nearly optimal dependence on all parameters. In 2015 IEEE 56th annual symposium on foundations of computer science, pages 792–809. IEEE, 2015

    2015

  8. [16]

    On the relationship between continuous-and discrete-time quantum walk

    Andrew M Childs. On the relationship between continuous-and discrete-time quantum walk. Communications in Mathe- matical Physics, 294(2):581–603, 2010

    2010

  9. [17]

    Optimal hamiltonian simulation by quantum signal processing

    Guang Hao Low and Isaac L Chuang. Optimal hamiltonian simulation by quantum signal processing. Physical review letters, 118(1):010501, 2017

    2017

  10. [18]

    Hamiltonian simulation by qubitization

    Guang Hao Low and Isaac L Chuang. Hamiltonian simulation by qubitization. Quantum, 3:163, 2019

    2019

  11. [19]

    Time-dependent hamiltonian simulation with l1-norm scaling

    Dominic W Berry, Andrew M Childs, Yuan Su, Xin Wang, and Nathan Wiebe. Time-dependent hamiltonian simulation with l1-norm scaling. Quantum, 4:254, 2020

    2020

  12. [20]

    Time-dependent hamiltonian simulation using discrete- clock constructions

    Jacob Watkins, Nathan Wiebe, Alessandro Roggero, and Dean Lee. Time-dependent hamiltonian simulation using discrete- clock constructions. PRX Quantum, 5(4):040316, 2024

    2024

  13. [21]

    Quantum algorithm for time-dependent hamiltonian simulation by permutation expansion

    Yi-Hsiang Chen, Amir Kalev, and Itay Hen. Quantum algorithm for time-dependent hamiltonian simulation by permutation expansion. PRX Quantum, 2(3):030342, 2021

    2021

  14. [22]

    Time-dependent hamiltonian simulation of highly oscillatory dynamics and superconver- gence for schr¨ odinger equation.Quantum, 6:690, 2022

    Dong An, Di Fang, and Lin Lin. Time-dependent hamiltonian simulation of highly oscillatory dynamics and superconver- gence for schr¨ odinger equation.Quantum, 6:690, 2022

    2022

  15. [23]

    Time-dependent unbounded hamiltonian simulation with vector norm scaling

    Dong An, Di Fang, and Lin Lin. Time-dependent unbounded hamiltonian simulation with vector norm scaling. Quantum, 5:459, 2021

    2021

  16. [24]

    Simulating the dynamics of time-dependent hamiltonians with a truncated dyson series

    M´ aria Kieferov´ a, Artur Scherer, and Dominic W Berry. Simulating the dynamics of time-dependent hamiltonians with a truncated dyson series. Physical Review A, 99(4):042314, 2019

    2019

  17. [25]

    Hamiltonian simulation in the interaction picture

    Guang Hao Low and Nathan Wiebe. Hamiltonian simulation in the interaction picture. arXiv preprint arXiv:1805.00675, 2018

    2018 arXiv

  18. [26]

    Quantum walk algorithm for element distinctness

    Andris Ambainis. Quantum walk algorithm for element distinctness. SIAM Journal on Computing, 37(1):210–239, 2007

    2007

  19. [27]

    Quantum algorithm for linear systems of equations

    Aram W Harrow, Avinatan Hassidim, and Seth Lloyd. Quantum algorithm for linear systems of equations. Physical review letters, 103(15):150502, 2009

    2009

  20. [28]

    Quantum algorithm for systems of linear equations with exponentially improved dependence on precision

    Andrew M Childs, Robin Kothari, and Rolando D Somma. Quantum algorithm for systems of linear equations with exponentially improved dependence on precision. SIAM Journal on Computing, 46(6):1920–1950, 2017

    1920

  21. [29]

    Quantum algorithm for data fitting

    Nathan Wiebe, Daniel Braun, and Seth Lloyd. Quantum algorithm for data fitting. Physical review letters, 109(5):050505, 2012

    2012

  22. [30]

    A quantum algorithm to solve nonlinear differential equations

    Sarah K Leyton and Tobias J Osborne. A quantum algorithm to solve nonlinear differential equations. arXiv preprint arXiv:0812.4423, 2008. 11

    2008 arXiv

  23. [31]

    Efficient quantum algorithm for dissipative nonlinear differential equations

    Jin-Peng Liu, Herman Øie Kolden, Hari K Krovi, Nuno F Loureiro, Konstantina Trivisa, and Andrew M Childs. Efficient quantum algorithm for dissipative nonlinear differential equations. Proceedings of the National Academy of Sciences, 118(35):e2026805118, 2021

    2021

  24. [32]

    Quantum algorithm for linear differential equations with exponentially improved dependence on precision

    Dominic W Berry, Andrew M Childs, Aaron Ostrander, and Guoming Wang. Quantum algorithm for linear differential equations with exponentially improved dependence on precision. Communications in Mathematical Physics, 356:1057–1081, 2017

    2017

  25. [33]

    Quantum computing for nonlinear differential equations and turbulence

    Felix Tennie, Sylvain Laizet, Seth Lloyd, and Luca Magri. Quantum computing for nonlinear differential equations and turbulence. arXiv preprint arXiv:2406.04826, 2024

    2024 arXiv

  26. [34]

    Quantum circuit learning

    Kosuke Mitarai, Makoto Negoro, Masahiro Kitagawa, and Keisuke Fujii. Quantum circuit learning. Physical Review A, 98(3):032309, 2018

    2018

  27. [35]

    The quest for a quantum neural network

    Maria Schuld, Ilya Sinayskiy, and Francesco Petruccione. The quest for a quantum neural network. Quantum Information Processing, 13(11):2567–2586, 2014

    2014

  28. [36]

    Supervised learning with quantum computers, volume 17

    Maria Schuld and Francesco Petruccione. Supervised learning with quantum computers, volume 17. Springer, 2018

    2018

  29. [37]

    Evaluating analytic gradients on quantum hardware

    Maria Schuld, Ville Bergholm, Christian Gogolin, Josh Izaac, and Nathan Killoran. Evaluating analytic gradients on quantum hardware. Physical Review A, 99(3):032331, 2019

    2019

  30. [38]

    Machine learning in quantum spaces, 2019

    Maria Schuld. Machine learning in quantum spaces, 2019

    2019

  31. [39]

    Quantum machine learning in feature hilbert spaces

    Maria Schuld and Nathan Killoran. Quantum machine learning in feature hilbert spaces. Physical review letters, 122(4):040504, 2019

    2019

  32. [40]

    Circuit-centric quantum classifiers

    Maria Schuld, Alex Bocharov, Krysta M Svore, and Nathan Wiebe. Circuit-centric quantum classifiers. Physical Review A, 101(3):032308, 2020

    2020

  33. [41]

    Supervised learning with quantum-enhanced feature spaces

    Vojtˇ ech Havl ´ ıˇ cek, Antonio D C´ orcoles, Kristan Temme, Aram W Harrow, Abhinav Kandala, Jerry M Chow, and Jay M Gambetta. Supervised learning with quantum-enhanced feature spaces. Nature, 567(7747):209–212, 2019

    2019

  34. [42]

    A rigorous and robust quantum speed-up in supervised machine learning

    Yunchao Liu, Srinivasan Arunachalam, and Kristan Temme. A rigorous and robust quantum speed-up in supervised machine learning. Nature Physics, 17(9):1013–1017, 2021

    2021

  35. [43]

    Quantum singular value transformation and beyond: exponential improvements for quantum matrix arithmetics

    Andr´ as Gily´ en, Yuan Su, Guang Hao Low, and Nathan Wiebe. Quantum singular value transformation and beyond: exponential improvements for quantum matrix arithmetics. In Proceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing, pages 193–204, 2019

    2019

  36. [44]

    Quantum algorithm for petz recovery channels and pretty good measurements

    Andr´ as Gily´ en, Seth Lloyd, Iman Marvian, Yihui Quek, and Mark M Wilde. Quantum algorithm for petz recovery channels and pretty good measurements. Physical Review Letters, 128(22):220502, 2022

    2022

  37. [45]

    Principal component analysis without centering

    uses a trick to reduce the eigenvalue finding to solving a small linear system size, thus avoiding the measurement step, which possesses an exponential barrier. On the other hand, the method introduced by the same authors in [50] utilizes the recently introduced quantum singul...

  38. [46]

    Improved quantum algorithms for eigenvalues finding and gradient descent

    Nhat A Nghiem and Tzu-Chieh Wei. Improved quantum algorithms for eigenvalues finding and gradient descent. arXiv preprint arXiv:2312.14786, 2023

    2023 arXiv

  39. [47]

    Quantum gradient descent and newton’s method for constrained polynomial optimization

    Patrick Rebentrost, Maria Schuld, Leonard Wossnig, Francesco Petruccione, and Seth Lloyd. Quantum gradient descent and newton’s method for constrained polynomial optimization. New Journal of Physics, 21(7):073023, 2019

    2019

  40. [48]

    Perturbation theory with quantum signal processing

    Kosuke Mitarai, Kiichiro Toyoizumi, and Wataru Mizukami. Perturbation theory with quantum signal processing. Quan- tum, 7:1000, 2023

    2023

  41. [49]

    Quantum amplitude amplification and estimation

    Gilles Brassard, Peter Hoyer, Michele Mosca, and Alain Tapp. Quantum amplitude amplification and estimation. Con- temporary Mathematics, 305:53–74, 2002

    2002

  42. [50]

    Quantum algorithm for estimating largest eigenvalues

    Nhat A Nghiem and Tzu-Chieh Wei. Quantum algorithm for estimating largest eigenvalues. Physics Letters A, 488:129138, 2023

    2023

  43. [51]

    Improved quantum power method and numerical integration using quantum singular value transformation

    Nhat A Nghiem, Hiroki Sukeno, Shuyu Zhang, and Tzu-Chieh Wei. Improved quantum power method and numerical integration using quantum singular value transformation. arXiv preprint arXiv:2407.11744, 2024

    2024 arXiv

  44. [52]

    Synthesis of quantum superpositions by quantum computation

    Lov K Grover. Synthesis of quantum superpositions by quantum computation. Physical review letters, 85(6):1334, 2000

    2000

  45. [53]

    Creating superpositions that correspond to efficiently integrable probability distributions

    Lov Grover and Terry Rudolph. Creating superpositions that correspond to efficiently integrable probability distributions. arXiv preprint quant-ph/0208112, 2002

    2002 arXiv

  46. [54]

    Quantum-state preparation with universal gate decompositions

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

    2011

  47. [55]

    Approximate amplitude encoding in shallow parameterized quantum circuits and its application to financial market indicators

    Kouhei Nakaji, Shumpei Uno, Yohichi Suzuki, Rudy Raymond, Tamiya Onodera, Tomoki Tanaka, Hiroyuki Tezuka, Naoki Mitsuda, and Naoki Yamamoto. Approximate amplitude encoding in shallow parameterized quantum circuits and its application to financial market indicators. Physical Re...

    2022

  48. [56]

    Quantum algorithms for approximate function loading

    Gabriel Marin-Sanchez, Javier Gonzalez-Conde, and Mikel Sanz. Quantum algorithms for approximate function loading. Physical Review Research, 5(3):033114, 2023

    2023

  49. [57]

    An improved method for quantum matrix multiplication

    Nhat A Nghiem and Tzu-Chieh Wei. An improved method for quantum matrix multiplication. Quantum Information Processing, 22(8):299, 2023

    2023

  50. [58]

    A quantum-inspired classical algorithm for recommendation systems

    Ewin Tang. A quantum-inspired classical algorithm for recommendation systems. In Proceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing, pages 217–228, 2019

    2019

  51. [59]

    Quantum principal component analysis only achieves an exponential speedup because of its state preparation assumptions

    Ewin Tang. Quantum principal component analysis only achieves an exponential speedup because of its state preparation assumptions. Physical Review Letters, 127(6):060503, 2021

    2021

  52. [60]

    Quantum-inspired low-rank stochastic regression with logarithmic dependence on the dimension

    Andr´ as Gily´ en, Seth Lloyd, and Ewin Tang. Quantum-inspired low-rank stochastic regression with logarithmic dependence on the dimension. arXiv preprint arXiv:1811.04909, 2018

    2018 arXiv

  53. [61]

    Approximate quantum circuit synthesis using block encodings

    Daan Camps and Roel Van Beeumen. Approximate quantum circuit synthesis using block encodings. Physical Review A, 102(5):052411, 2020

    2020

  54. [62]

    Lecture notes on quantum algorithms

    Andrew M Childs. Lecture notes on quantum algorithms. Lecture notes at University of Maryland, 2017. 12 PRELIMINARIES Here, we summarize the main recipes of our work, which mostly derived in the seminal QSVT work [43]. We keep the statements brief and precise for simplicity, w...

    2017

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.