pith. machine review for the scientific record. sign in

arxiv: 2603.24654 · v2 · submitted 2026-03-25 · 🪐 quant-ph · cs.LG· stat.ML

Recognition: unknown

Spectral methods: crucial for machine learning, natural for quantum computers?

Vasilis Belis , Joseph Bowles , Rishabh Gupta , Evan Peters , Maria Schuld
Authors on Pith no claims yet

Pith reviewed 2026-05-15 00:19 UTC · model grok-4.3

classification 🪐 quant-ph cs.LGstat.ML
keywords quantum machine learningspectral methodsFourier spectrumQuantum Fourier Transformgenerative modelsspectral biasdeep learningregularisation
0
0 comments X

The pith

Quantum computers can manipulate the Fourier spectrum of machine learning models more directly than classical methods via the Quantum Fourier Transform.

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

The paper argues that spectral methods, which learn, regularise or manipulate the Fourier spectrum of a model, are both fundamental to machine learning and naturally suited to quantum computers. Representing a generative model as a quantum state lets the Quantum Fourier Transform access and alter its spectrum using the full set of quantum operations, something usually prohibitive on classical hardware. This matters because a spectral bias may explain why deep networks succeed, support vector machines have long regularised in Fourier space, and convolutional nets build their filters there too. If the mapping holds, quantum computers could enable fundamentally different and more resource-efficient control over these spectral properties. The authors present the case to encourage quantum machine learning work that starts from the question of why a quantum approach is advantageous.

Core claim

If a generative machine learning model is represented by a quantum state, the Quantum Fourier Transform allows us to manipulate the Fourier spectrum of the state using the entire toolbox of quantum routines, an operation that is usually prohibitive for classical models. Spectral methods are surprisingly fundamental to machine learning: a spectral bias has been hypothesised as the core principle behind deep learning success, support vector machines regularise in Fourier space, and convolutional neural nets build filters in the Fourier space of images.

What carries the argument

The Quantum Fourier Transform applied to quantum states that represent machine learning models, granting direct access to spectral manipulation through quantum routines.

If this is right

  • Generative models could have their spectra engineered directly on quantum hardware rather than through costly classical transforms.
  • Regularisation techniques used in support vector machines could be implemented more naturally in Fourier space on quantum devices.
  • The spectral bias observed in deep networks might be controlled or tested by direct quantum operations on model states.
  • Convolutional filters in image models could be constructed using quantum Fourier routines instead of classical convolution.

Where Pith is reading between the lines

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

  • Quantum advantage claims in machine learning could be sharpened by focusing on tasks that require precise, high-dimensional Fourier control.
  • Near-term devices might demonstrate the approach on small generative models by measuring spectral properties after applying the Quantum Fourier Transform.
  • Hybrid workflows could combine classical training with quantum spectral regularisation steps where classical Fourier analysis scales poorly.

Load-bearing premise

Representing an ML model as a quantum state and applying quantum Fourier routines will yield fundamentally more direct or resource-efficient spectral control than classical methods without prohibitive overheads in state preparation or measurement.

What would settle it

Showing that a classical algorithm can achieve equivalent Fourier-spectrum manipulation for comparable generative models with equal or better efficiency and without quantum state overhead would falsify the proposed advantage.

Figures

Figures reproduced from arXiv: 2603.24654 by Evan Peters, Joseph Bowles, Maria Schuld, Rishabh Gupta, Vasilis Belis.

Figure 1
Figure 1. Figure 1: FIG. 1. A common simplicity bias of machine learning models [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. A sketch of the idea of smoothing an empirical dis [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. An example of smoothing an empirical distribution in Fourier space to solve the generative learning problem. Starting [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. An example of smoothing empirical distributions with quantum computers, using the same filter as in Figure [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Illustration of the relation between a group and a [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Fourier basis functions [PITH_FULL_IMAGE:figures/full_fig_p012_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8. Examples for basis functions ˆm [PITH_FULL_IMAGE:figures/full_fig_p014_8.png] view at source ↗
read the original abstract

This article presents an argument for why quantum computers could unlock new methods for machine learning. We argue that spectral methods, in particular those that learn, regularise, or otherwise manipulate the Fourier spectrum of a machine learning model, are often natural for quantum computers. For example, if a generative machine learning model is represented by a quantum state, the Quantum Fourier Transform allows us to manipulate the Fourier spectrum of the state using the entire toolbox of quantum routines, an operation that is usually prohibitive for classical models. At the same time, spectral methods are surprisingly fundamental to machine learning: A spectral bias has recently been hypothesised to be the core principle behind the success of deep learning; support vector machines have been known for decades to regularise in Fourier space, and convolutional neural nets build filters in the Fourier space of images. Could, then, quantum computing open fundamentally different, much more direct and resource-efficient ways to design the spectral properties of a model? We discuss this potential in detail here, hoping to stimulate a direction in quantum machine learning research that puts the question of ``why quantum?'' first.

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

1 major / 2 minor

Summary. The paper presents a conceptual argument that spectral methods in machine learning—particularly those that learn, regularize, or manipulate the Fourier spectrum of a model—are naturally suited to quantum computers. It claims that representing a generative ML model as a quantum state allows the Quantum Fourier Transform (QFT) and associated quantum routines to directly manipulate its Fourier spectrum, an operation that is usually prohibitive classically. The manuscript connects this to the hypothesized spectral bias underlying deep learning success, Fourier-space regularization in support vector machines, and filter construction in convolutional neural networks, and advocates for quantum ML research that prioritizes the question of 'why quantum?' through spectral methods.

Significance. If the efficiency claims can be substantiated with concrete protocols, this perspective could open a productive research direction in quantum machine learning by identifying spectral manipulation as a core, resource-efficient advantage of quantum hardware. It builds on established ML observations (spectral bias, SVM regularization) and quantum primitives (QFT) to argue for fundamentally different model design approaches, which could stimulate falsifiable comparisons between quantum and classical spectral methods.

major comments (1)
  1. [Abstract] Abstract: the central efficiency claim—that QFT-based spectral manipulation on a quantum-state representation of an ML model is 'much more direct and resource-efficient' than classical methods—is load-bearing but unsupported. No concrete state-preparation protocol, circuit depth or qubit count, measurement overhead analysis, or comparison to classical Fourier methods (e.g., FFT complexity for kernel regularization) is supplied, leaving open the possibility that preparation and readout costs dominate for any non-trivial task.
minor comments (2)
  1. The manuscript would benefit from a dedicated subsection that sketches at least one explicit encoding (e.g., amplitude or phase encoding of a kernel or weight vector) and the corresponding QFT circuit, even at a high level, to make the 'natural for quantum' claim more operational.
  2. References to 'the entire toolbox of quantum routines' for spectral manipulation should be accompanied by one or two concrete examples (e.g., phase estimation or variational circuits acting on the Fourier basis) to avoid remaining at the level of analogy.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their thoughtful review and constructive feedback. We appreciate the acknowledgment of the manuscript's potential to open a productive research direction. We respond to the major comment as follows.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central efficiency claim—that QFT-based spectral manipulation on a quantum-state representation of an ML model is 'much more direct and resource-efficient' than classical methods—is load-bearing but unsupported. No concrete state-preparation protocol, circuit depth or qubit count, measurement overhead analysis, or comparison to classical Fourier methods (e.g., FFT complexity for kernel regularization) is supplied, leaving open the possibility that preparation and readout costs dominate for any non-trivial task.

    Authors: We agree that the paper is a conceptual perspective and does not provide concrete protocols or complexity analyses for state preparation, circuit depths, or comparisons to classical methods such as the FFT. The efficiency claim is meant to highlight a potential advantage based on the direct applicability of the QFT to quantum states representing models, but we recognize that preparation and measurement overheads must be considered and could be dominant. To address this valid point, we will revise the abstract to clarify that this is a hypothesized advantage motivating further research, rather than an established fact. We will also add text in the manuscript discussing the challenges associated with state preparation in this context. revision: yes

Circularity Check

0 steps flagged

No circularity: perspective argument relies on external established concepts

full rationale

The manuscript is a perspective piece without new derivations, equations, or quantitative predictions. It invokes known ML phenomena (spectral bias in deep learning, Fourier regularization in SVMs, CNN filters) and the standard QFT as external facts. No step reduces a claimed result to a fitted parameter, self-citation chain, or renamed input by construction. The central thesis remains a conceptual suggestion rather than a closed logical loop.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Review is based solely on the abstract; no explicit free parameters, new entities, or detailed axioms are introduced. The core assumption is that quantum spectral operations align naturally with ML needs.

axioms (1)
  • domain assumption Spectral methods are fundamental to the success of deep learning and other ML models
    Invoked via references to spectral bias hypothesis, SVM regularization, and CNN filters.

pith-pipeline@v0.9.0 · 5501 in / 1172 out tokens · 44144 ms · 2026-05-15T00:19:48.096615+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. Toward Generative Quantum Utility via Correlation-Complexity Map

    cs.LG 2026-03 unverdicted novelty 7.0

    A pre-training diagnostic map based on spectral correlation resemblance to IQP circuits and excess structural complexity identifies suitable datasets like turbulence data for quantum generative models, yielding compet...

Reference graph

Works this paper leans on

125 extracted references · 125 canonical work pages · cited by 1 Pith paper · 10 internal anchors

  1. [1]

    super- polynomial

    Continuous features The standardcontinuous Fourier transform, which we mentioned in the introduction, transforms functions on the groupR(or more precisely, (R,+)), with characters exp (2πixk). Its multidimensional version maps functions on products of this group,R N. The Fourier coefficients of a function overR N are related to itssmoothness. The most obv...

    work page

  2. [2]

    spin variable

    Binary features In Section II we saw theWalsh transformthat acts on functions on a product of cyclic groups with two ele- ments,Z 2 × · · · ×Z2 =Z n 2 . This group can be thought of as the set of bitstringsx∈ {0,1} n with addition modulo 2, and gives rise to boolean cube arithmetic. We also saw that the Fourier coefficients of a probability distribution o...

    work page

  3. [3]

    objectsADEare in positions2,4,5, while objectsBC are in positions1,3,

    Permutations To venture into the realm of non-Abelian groups, we want to summarise the intuition that Persi Diaconis built for the notion of simplicity contained in the Fourier spec- trum of functions over the symmetric groupS n [23] (see also [65] for a gentle introduction). Loosely speaking, the Fourier coefficients are related to expected patterns such...

    work page

  4. [4]

    support vectors

    Stationary kernels are filters in Fourier space First, let us define what a kernel is. Definition 3.LetGbe a locally compact group. A kernel is a symmetric, positive definite functionκ:G×G→C. If a kernel only depends on the relative valuegg ′, it is known as a stationary kernel. Definition 4.A left stationary kernel is a kernel that is invariant with resp...

    work page

  5. [5]

    feature collapse

    propose to use a kernel whose Fourier spectrum is a normalised mixture of a finite number of Gaussian distri- butions with trainable means and deviations, and thereby designs the spectral bias directly. After deep learning be- came popular, kernel learning moved to the idea of using deep neural networks to extract the feature vectors fed into a kernel met...

    work page

  6. [6]

    F-Principle

    The spectral bias of deep learning Thespectral biasdescribes the tendency of deep neural networks to learn low-frequency components of a target function significantly faster than high-frequency compo- nents. The phenomenon was first observed concurrently by [32] and [33] (who used the term “F-Principle”). Ra- hamanet al.[32] use an analytic derivation tha...

    work page

  7. [7]

    rep- resents

    on the other hand compare averages of the low and high-order parts of the training error’s Fourier spectrum, with similar conclusions. The computational method to make statements about the Fourier transform of high- dimensional functions was later refined by Kiessling and Thor [81], who use a sinc-kernel to extract the averages by convolution—once more ma...

    work page

  8. [8]

    quantum neural networks

    (rather than computing it from an oracle), although it is yet unclear how this problem relates to real-world applications. D. Spectral methods beyond the QFT The situation of spectral biases looks very different for “quantum neural networks” [10, 35]. Here the computa- tional basis that the QFT acts on is not associated with the data any more, which is in...

    work page

  9. [9]

    Born encoding

    proposed operating in a bandlimited Fourier space, which entails retaining only the low-order Fourier coeffi- cients. Such models can represent simple correlations in the data, and solve specific inference tasks that, likewise, only require these low-order Fourier coefficients. How- ever, even with advanced implementations and severe bandlimiting, these s...

    work page

  10. [10]

    resource fingerprint

    (see also the PennyLane demo [103]). Resource the- ories in quantum information ask how “complex” a given quantum state is with respect to a certain measure of complexity, which often translates into how difficult they are to prepare in the lab, or how difficult they are to simulate on a classical computer. Examples of resources areentanglement,Clifford s...

    work page

  11. [11]

    Babbush, R

    R. Babbush, R. King, S. Boixo, W. Huggins, T. Khat- tar, G. H. Low, J. R. McClean, T. O’Brien, and N. C. Rubin, The grand challenge of quantum applications 23 (2025), arXiv:2511.09124

    work page arXiv 2025

  12. [12]

    Schuld and N

    M. Schuld and N. Killoran, Is quantum advantage the right goal for quantum machine learning?, PRX Quan- tum3, 030101 (2022)

    work page 2022

  13. [13]

    Cerezo, G

    M. Cerezo, G. Verdon, H.-Y. Huang, L. Cincio, and P. J. Coles, Challenges and opportunities in quantum machine learning, Nature computational science2, 567 (2022)

    work page 2022

  14. [14]

    Schuld and F

    M. Schuld and F. Petruccione,Machine learning with quantum computers, Vol. 676 (Springer, 2021)

    work page 2021

  15. [15]

    Biamonte, P

    J. Biamonte, P. Wittek, N. Pancotti, P. Rebentrost, N. Wiebe, and S. Lloyd, Quantum machine learning, Nature549, 195 (2017)

    work page 2017

  16. [16]

    Rebentrost, M

    P. Rebentrost, M. Mohseni, and S. Lloyd, Quantum sup- port vector machine for big data classification, Physical review letters113, 130503 (2014)

    work page 2014

  17. [17]

    Quantum Recommendation Systems

    I. Kerenidis and A. Prakash, Quantum recommendation systems (2016), arXiv:1603.08675

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  18. [18]

    Wiebe, D

    N. Wiebe, D. Braun, and S. Lloyd, Quantum algorithm for data fitting, Physical review letters109, 050505 (2012)

    work page 2012

  19. [19]

    Lloyd, M

    S. Lloyd, M. Mohseni, and P. Rebentrost, Quantum principal component analysis, Nature Physics10, 631 (2014)

    work page 2014

  20. [20]

    Classification with Quantum Neural Networks on Near Term Processors

    E. Farhi and H. Neven, Classification with quan- tum neural networks on near term processors (2018), arXiv:1802.06002

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  21. [21]

    Schuld, A

    M. Schuld, A. Bocharov, K. M. Svore, and N. Wiebe, Circuit-centric quantum classifiers, Physical Review A 101, 032308 (2020)

    work page 2020

  22. [22]

    Cerezo, M

    M. Cerezo, M. Larocca, D. Garc´ ıa-Mart´ ın, N. L. Diaz, P. Braccia, E. Fontana, M. S. Rudolph, P. Bermejo, A. Ijaz, S. Thanasilp,et al., Does provable absence of barren plateaus imply classical simulability?, Nature Communications16, 7907 (2025)

    work page 2025

  23. [23]

    Bowles, S

    J. Bowles, S. Ahmed, and M. Schuld, Better than clas- sical? the subtle art of benchmarking quantum machine learning models (2024), arXiv:2403.07059

    work page arXiv 2024

  24. [24]

    Recio-Armengol, S

    E. Recio-Armengol, S. Ahmed, and J. Bowles, Train on classical, deploy on quantum: scaling generative quantum machine learning to a thousand qubits (2025), arXiv:2503.02934

    work page arXiv 2025

  25. [25]

    Huang, M

    H.-Y. Huang, M. Broughton, N. Eassa, H. Neven, R. Babbush, and J. R. McClean, Generative quantum advantage for classical and quantum problems (2025), arXiv:2509.09033

    work page arXiv 2025

  26. [26]

    Huang, M

    H.-Y. Huang, M. Broughton, J. Cotler, S. Chen, J. Li, M. Mohseni, H. Neven, R. Babbush, R. Kueng, J. Preskill,et al., Quantum advantage in learning from experiments, Science376, 1182 (2022)

    work page 2022

  27. [27]

    A. G. Wilson, Deep learning is not so mysterious or different (2025), arXiv:2503.02113

    work page arXiv 2025

  28. [28]

    Srivastava, G

    N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhutdinov, Dropout: A simple way to pre- vent neural networks from overfitting, Journal of Ma- chine Learning Research15, 1929 (2014)

    work page 1929

  29. [29]

    H. Noh, T. You, J. Mun, and B. Han, Regularizing deep neural networks by noise: Its interpretation and opti- mization, inAdvances in Neural Information Processing Systems, Vol. 30 (2017)

    work page 2017

  30. [30]

    Miyato, T

    T. Miyato, T. Kataoka, M. Koyama, and Y. Yoshida, Spectral normalization for generative adversarial net- works, inInternational Conference on Learning Repre- sentations(2018)

    work page 2018

  31. [31]

    Dherin, M

    B. Dherin, M. Munn, M. Rosca, and D. G. Barrett, Why neural networks find simple solutions: The many regu- larizers of geometric complexity, inAdvances in Neural Information Processing Systems, edited by A. H. Oh, A. Agarwal, D. Belgrave, and K. Cho (2022)

    work page 2022

  32. [32]

    Rosca, T

    M. Rosca, T. Weber, A. Gretton, and S. Mohamed, A case for new neural network smoothness constraints, inProceedings on ”I Can’t Believe It’s Not Better!” at NeurIPS Workshops, Proceedings of Machine Learning Research, Vol. 137, edited by J. Zosa Forde, F. Ruiz, M. F. Pradier, and A. Schein (PMLR, 2020) pp. 21–32

    work page 2020

  33. [33]

    Diaconis, Group representations in probability and statistics, Lecture Notes-Monograph Series11, i (1988)

    P. Diaconis, Group representations in probability and statistics, Lecture Notes-Monograph Series11, i (1988)

    work page 1988

  34. [34]

    I. R. Kondor,Group theoretical methods in machine learning(Columbia University, 2008)

    work page 2008

  35. [35]

    Sch¨ olkopf and A

    B. Sch¨ olkopf and A. J. Smola,Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond(The MIT Press, 2018)

    work page 2018

  36. [36]

    Steinwart and A

    I. Steinwart and A. Christmann,Support vector ma- chines(Springer Science & Business Media, 2008)

    work page 2008

  37. [37]

    Gretton, K

    A. Gretton, K. M. Borgwardt, M. J. Rasch, B. Sch¨ olkopf, and A. Smola, A kernel two-sample test, The Journal of Machine Learning Research13, 723 (2012)

    work page 2012

  38. [38]

    Li, W.-C

    C.-L. Li, W.-C. Chang, Y. Cheng, Y. Yang, and B. P´ oczos, Mmd gan: Towards deeper understanding of moment matching network, Advances in Neural In- formation Processing Systems30(2017)

    work page 2017

  39. [39]

    M. M. Bronstein, J. Bruna, T. Cohen, and P. Veliˇ ckovi´ c, Geometric deep learning: Grids, groups, graphs, geodesics, and gauges (2021), arXiv:2104.13478

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  40. [40]

    Kondor and S

    R. Kondor and S. Trivedi, On the generalization of equivariance and convolution in neural networks to the action of compact groups, inInternational Conference on Machine Learning(PMLR, 2018) pp. 2747–2755

    work page 2018

  41. [41]

    Z. Wu, S. Pan, F. Chen, G. Long, C. Zhang, and P. S. Yu, A comprehensive survey on graph neural networks, IEEE transactions on neural networks and learning sys- tems32, 4 (2020)

    work page 2020

  42. [42]

    Rahaman, A

    N. Rahaman, A. Baratin, D. Arpit, F. Draxler, M. Lin, F. Hamprecht, Y. Bengio, and A. Courville, On the spectral bias of neural networks, inInternational Con- ference on Machine Learning(PMLR, 2019) pp. 5301– 5310

    work page 2019

  43. [43]

    Z.-Q. J. Xu, Y. Zhang, T. Luo, Y. Xiao, and Z. Ma, Fre- quency principle: Fourier analysis sheds light on deep neural networks (2019), arXiv:1901.06523

    work page arXiv 2019

  44. [44]

    Mitarai, M

    K. Mitarai, M. Negoro, M. Kitagawa, and K. Fu- jii, Quantum circuit learning, Physical Review A98, 032309 (2018)

    work page 2018

  45. [45]

    Schuld, V

    M. Schuld, V. Bergholm, C. Gogolin, J. Izaac, and N. Killoran, Evaluating analytic gradients on quantum hardware, Physical Review A99, 032331 (2019)

    work page 2019

  46. [46]

    Liu and L

    J.-G. Liu and L. Wang, Differentiable learning of quan- tum circuit born machines, Physical Review A98, 062324 (2018)

    work page 2018

  47. [47]

    M. S. Rudolph, S. Lerch, S. Thanasilp, O. Kiss, O. Shaya, S. Vallecorsa, M. Grossi, and Z. Holmes, Trainability barriers and opportunities in quantum gen- erative modeling, npj Quantum Information10, 116 (2024)

    work page 2024

  48. [48]

    Moore, D

    C. Moore, D. Rockmore, and A. Russell, Generic quan- tum fourier transforms, ACM Transactions on Algo- rithms (TALG)2, 707 (2006). 24

    work page 2006

  49. [49]

    Schuld, R

    M. Schuld, R. Sweke, and J. J. Meyer, Effect of data encoding on the expressive power of variational quantum-machine-learning models, Physical Review A 103, 032430 (2021)

    work page 2021

  50. [50]

    R. Lu, R. Zhang, W. Li, Z. Wei, D.-L. Deng, and Z. Liu, A unified frequency principle for quantum and classical machine learning (2026), arXiv:2601.03169

    work page arXiv 2026

  51. [51]

    Duffy and M

    C. Duffy and M. Jastrzebski, Spectral bias in variational quantum machine learning (2025), arXiv:2506.22555

    work page arXiv 2025

  52. [52]

    The spectral amplitude principle for dynamics of quantum neural networks,

    Y.-h. Xu, D.-B. Zhang, and J. Yan, The spectral ampli- tude principle for dynamics of quantum neural networks (2025), arXiv:2409.06682

    work page arXiv 2025

  53. [53]

    Sweke, S

    R. Sweke, S. Shin, and E. Gil-Fuster, Kernel-based de- quantization of variational qml without random fourier features (2025), arXiv:2503.23931

    work page arXiv 2025

  54. [54]

    Sweke, E

    R. Sweke, E. Recio-Armengol, S. Jerbi, E. Gil-Fuster, B. Fuller, J. Eisert, and J. J. Meyer, Potential and limitations of random fourier features for dequantizing quantum machine learning, Quantum9, 1640 (2025)

    work page 2025

  55. [55]

    A. M. Childs and W. van Dam, Quantum algorithms for algebraic problems, Reviews of Modern Physics82, 1 (2010)

    work page 2010

  56. [56]

    Bouland, T

    A. Bouland, T. Giurgica-Tiron, and J. Wright, The state hidden subgroup problem and an efficient algorithm for locating unentanglement (2024), arXiv:2410.12706

    work page arXiv 2024

  57. [57]

    Simidzija, E

    P. Simidzija, E. Koskin, E. Y. Zhu, M. Dascal, and M. Schuld, Solving approximate hidden subgroup prob- lems: quantum heuristics to detect weak entanglement (2026), arXiv:2603.15733 [quant-ph]

    work page arXiv 2026

  58. [58]

    N. Guo, K. Mitarai, and K. Fujii, Nonlinear transforma- tion of complex amplitudes via quantum singular value transformation, Physical Review Research6, 043227 (2024)

    work page 2024

  59. [59]

    A. G. Rattew and P. Rebentrost, Non-linear trans- formations of quantum amplitudes: Exponential im- provement, generalization, and applications (2023), arXiv:2309.09839

    work page arXiv 2023

  60. [60]

    A. Y. Kitaev, Quantum measurements and the abelian stabilizer problem (1995), arXiv:quant-ph/9511026

    work page internal anchor Pith review Pith/arXiv arXiv 1995

  61. [61]

    V. N. Vapnik,The nature of statistical learning theory (Springer-Verlag New York, Inc., 1995)

    work page 1995

  62. [62]

    L. G. Valiant, A theory of the learnable, Communica- tions of the ACM27, 1134 (1984)

    work page 1984

  63. [63]

    Vapnik and A

    V. Vapnik and A. Y. Chervonenkis, On the uniform con- vergence of relative frequencies of events to their proba- bilities, Theory of Probability and its Applications16, 264 (1971)

    work page 1971

  64. [64]

    Blumer, A

    A. Blumer, A. Ehrenfeucht, D. Haussler, and M. K. Warmuth, Learnability and the vapnik-chervonenkis di- mension, Journal of the ACM (JACM)36, 929 (1989)

    work page 1989

  65. [65]

    Hastie, R

    T. Hastie, R. Tibshirani, J. H. Friedman, and J. H. Friedman,The elements of statistical learning: data mining, inference, and prediction, Vol. 2 (Springer, 2009)

    work page 2009

  66. [66]

    P. L. Bartlett and S. Mendelson, Rademacher and gaus- sian complexities: Risk bounds and structural results, Journal of Machine Learning Research3, 463 (2002)

    work page 2002

  67. [67]

    Understanding deep learning requires rethinking generalization

    C. Zhang, S. Bengio, M. Hardt, B. Recht, and O. Vinyals, Understanding deep learning requires re- thinking generalization (2016), arXiv:1611.03530

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  68. [68]

    Belkin, D

    M. Belkin, D. Hsu, S. Ma, and S. Mandal, Reconcil- ing modern machine-learning practice and the classi- cal bias–variance trade-off, Proceedings of the National Academy of Sciences116, 15849 (2019)

    work page 2019

  69. [69]

    Jacot, F

    A. Jacot, F. Gabriel, and C. Hongler, Neural tangent kernel: Convergence and generalization in neural net- works, Advances in Neural Information Processing Sys- tems31(2018)

    work page 2018

  70. [70]

    Kalimeris, G

    D. Kalimeris, G. Kaplun, P. Nakkiran, B. Edelman, T. Yang, B. Barak, and H. Zhang, Sgd on neural net- works learns functions of increasing complexity, Ad- vances in Neural Information Processing Systems32 (2019)

    work page 2019

  71. [71]

    N. S. Keskar, D. Mudigere, J. Nocedal, M. Smelyanskiy, and P. T. P. Tang, On large-batch training for deep learning: Generalization gap and sharp minima (2016), arXiv:1609.04836

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  72. [72]

    The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks

    J. Frankle and M. Carbin, The lottery ticket hypoth- esis: Finding sparse, trainable neural networks (2018), arXiv:1803.03635

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  73. [73]

    R. E. Kirk, Experimental design, Sage handbook of quantitative methods in psychology , 23 (2009)

    work page 2009

  74. [74]

    R. A. Bailey, P. Diaconis, D. N. Rockmore, and C. Row- ley, A spectral analysis approach for experimental de- signs, Excursions in Harmonic Analysis4, 367 (2015)

    work page 2015

  75. [75]

    Huang, C

    J. Huang, C. Guestrin, and L. J. Guibas, Efficient in- ference for distributions on permutations, Advances in Neural Information Processing Systems20(2007)

    work page 2007

  76. [76]

    Huang, C

    J. Huang, C. Guestrin, and L. Guibas, Fourier theo- retic probabilistic inference over permutations., Journal of Machine Learning Research10(2009)

    work page 2009

  77. [77]

    Diaconis, The 1987 wald memorial lectures, The An- nals of Statistics17, 949 (1989)

    P. Diaconis, The 1987 wald memorial lectures, The An- nals of Statistics17, 949 (1989)

    work page 1987

  78. [78]

    B. L. Lawson, M. E. Orrison, and D. T. Uminsky, Spec- tral analysis of the supreme court, Mathematics Maga- zine79, 340 (2006)

    work page 2006

  79. [79]

    Belis, G

    V. Belis, G. Crognaletti, M. Argenton, M. Grossi, and M. Schuld, Probabilistic modeling over permutations using quantum computers (2026), arXiv:2603.22401 [quant-ph]

    work page arXiv 2026

  80. [80]

    Rahimi and B

    A. Rahimi and B. Recht, Random features for large- scale kernel machines, Advances in Neural Information Processing Systems20(2007)

    work page 2007

Showing first 80 references.