Pith. sign in

REVIEW 2 major objections 4 minor 60 references

Quantum channel learning with limited parallel access

T0 review · 2 major / 4 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read A master lower bound for quantum channel learning with limited parallel access shows tight sample-complexity scalings and a resource hierarchy across qudit and bosonic channels.

desk verdict The master lemma and the lower-bound hierarchy are serious progress, but the one upper bound that makes the easy side work has a genuine proof-sketch error and needs fixing before the ε^{-4} claim is settled. read the letter →

arxiv 2608.05307 v1 pith:74DGIFX4 submitted 2026-08-05 quant-ph

Mahadevan Subramanian , Hyukgun Kwon , Liang Jiang This is my paper · ORCID
classification quant-ph MSC 81P4581P5081P68
keywords quantumchannellearningsamplecomplexitytransfermatrixHeisenberg-Weyloperatorscomplex-conjugatemulti-copyprotocolsbosonicchannelsqudit
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 asks how many parallel uses of an unknown quantum channel are needed to learn its transfer matrix, and whether access to the complex-conjugate channel changes the answer. Its central claim is a master lower bound (Lemma III.1): any adaptive, ancilla-assisted protocol using $c$ copies per round requires depth $T=\Omega(1/\Delta)$ to solve a many-one channel-discrimination task, where $\Delta$ is a sum of operator norms over length-$c$ strings. From this single bound the authors derive matching upper and lower bounds: with access to $E\otimes E^*$ the sample complexity is $\Theta(\epsilon^{-4})$, without it a $d$-level system requires $d$ copies with tight $\epsilon^{-2d}$ scaling, and bosonic channels stay exponentially hard for all $c=O(1/\epsilon)$. The upshot is a hierarchy of channel-learning resources: one copy is hard for self-conjugate channels, two copies make them easy, and square-free dimensions need their square-free degree. A sympathetic reader should care because the learner controls inputs and ancillas, so these bounds are stronger than state-learning bounds and directly constrain practical noise characterization.

What carries the argument

The load-bearing object is Lemma III.1, the master lower bound. It bounds the maximal per-step $\chi^2$ divergence between two channel hypotheses by $\Delta$, defined in Eq. (19) as the squared sum over strings ${\bf x}\in\{0,\dots,l\}^c$ of the product of operator norms of averaged $c$-fold tensor powers of the operators $K$ and $W$ appearing in the channel's low-rank expansion. Any $c$-copy protocol with adaptive inputs, arbitrary ancillas, and arbitrary POVMs needs depth $T=\Omega(1/\Delta)$ for constant success probability above $1/2$; every later theorem is a specialization of this one to a specific family of channels, where the operator-norm transitions encode whether the relevant observables commute.

What would settle it

Numerically evaluate $Q(m/2,1)-Q(m/2,m/1.98)$ for every integer $8\leq m\leq 34000$; finding any value below $0.546$ would falsify the claimed bosonic lower bounds, while an analytic proof covering the whole range would close the last non-analytic gap.

Watch

Extended reading notes

Core claim

The paper's discovery is that the entire difficulty of learning a quantum channel under limited parallel access is captured by one master lemma, and that the transitions in difficulty are governed by whether certain tensor-power operators commute. For qudit channels of prime local dimension $d$, the operator-norm identity $\|\sum_{q,p}\hat D_{d,m}(q,p)^{\otimes 2k}\|_{\rm op}=d^m$ for $k\not\equiv 0\pmod d$ and $d^{2m}$ for $k\equiv 0\pmod d$ forces a sharp jump exactly at $c=d$ copies, giving the tight $\epsilon^{-2d}$ bound. With access to $E\otimes E^*$, the cross terms make the norm $\|\sum \hat D(q,p)^{\otimes 2k}\otimes \hat D(-q,p)^{\otimes 2l}\|$ saturate at $k\equiv l\pmod d$, which yields the tight $\epsilon^{-4}$ scaling. For multimode bosonic channels the analogous Gaussian-averaged displacement norms decay smoothly with the copy number, which is why no number $c=O(1/\epsilon)$ of copies escapes exponential hardness. The same lemma also yields sharper state-learning lower bounds as a corollary.

Load-bearing premise

The bosonic hardness results inherit a probability bound, Lemma B.7, that is verified manually by inspecting a figure for $8\leq m\leq 34000$ and proven only for $m\geq 34000$; if that constant is wrong, the bosonic lower bounds no longer hold.

Editorial extensions

If this is right

  • With simultaneous access to $E$ and $E^*$, a non-adaptive ancilla-assisted protocol achieves accuracy $\epsilon$ with $O(\log(M/\delta)\epsilon^{-4})$ samples, and the matching lower bound shows no $c$-copy protocol does better in the relevant regime.
  • Without the conjugate channel, any protocol using $c<d$ copies of a prime-$d$ qudit channel requires sample complexity exponential in the number of qudits, while $c=d$ recovers efficiency with tight $\epsilon^{-2d}$ scaling.
  • Self-conjugate channels are 1-copy hard but 2-copy easy, so the complex-conjugate resource cuts the required copy number from exponential to polynomial in one step.
  • For multimode bosonic channels, the sample complexity remains exponential for every $c=O(1/\epsilon)$ unless the conjugate channel is available, so the hierarchy is not flattened by any constant number of parallel copies.
  • Since channel learning with controlled inputs is at least as hard as Choi-state learning, the master lemma gives strictly stronger state-learning lower bounds with limited multi-copy access.

Reading between the lines

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

  • The master lemma should transfer to sequential strategies with bounded ancilla, but the paper leaves that case open; testing it would determine whether the $c$-copy hierarchy survives when copies are processed one at a time.
  • The precise value of the Gaussian probability constant in Lemma B.7 is worth an independent check: an analytic proof for $8\leq m\leq 34000$ would remove the one figure-based step from the bosonic theorems.
  • If query revelation were replaced by prior knowledge of the queries, the absolute-value-only restriction would be lifted and bosonic sign estimation might become efficient, shrinking the scope of the hardness claims; a concrete test would be to estimate phases of $C_E$ for Gaussian channels with one-copy access.
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

2 major / 4 minor

Summary. The paper studies the sample complexity of learning absolute values of transfer-matrix/characteristic-function entries of unknown quantum channels under limited parallel access. It introduces a general lower bound (Lemma III.1) for c-copy protocols with adaptive ancilla-assisted measurements, then specializes to qudit and bosonic channels: without complex-conjugate access, any c<d copies give exponential-in-system-size hardness, with tight ε^{-2d} scaling at c=d for prime d; with access to E⊗E*, the paper claims an O(ε^{-4} log(M/δ)) upper bound; and bosonic channels are claimed to be exponentially hard for all c=O(1/ε) without conjugate access. Matching lower bounds are intended to show the ε^{-4} and ε^{-2d} scalings are tight, and the paper draws a resource hierarchy in Fig. 4, including self-conjugate channels that are 1-copy hard but 2-copy easy. The proof of the central upper bound for the E⊗E* case, however, contains an accuracy-propagation error, and the bosonic lower bounds rest on a partially numerical Gaussian-concentration lemma.

Significance. If the results are established, Lemma III.1 is a substantial unifying tool for channel-learning lower bounds, extending prior state-learning hierarchies to the channel setting and providing tight ε-scalings. The paper is strong in its explicit operator-norm computations (Lemmas B.8, B.9, B.16, B.17), its careful treatment of infinite-dimensional POVMs, and its clear statement of the resource hierarchy (self-conjugate channels 1-copy hard but 2-copy easy; prime-d channels d-copy easy). The hierarchy is a clean, falsifiable set of claims, and the appendices are unusually detailed. However, the proof of Theorem IV.1, which is the only efficient upper bound for access to E⊗E*, currently contains an incorrect accuracy-propagation step, and the bosonic lower bounds in Theorems IV.3 and IV.5 inherit a finite-range numerical verification in Lemma B.7. Both issues are localized and repairable, but they must be fixed before the central conclusions can be regarded as proven.

major comments (2)
  1. [§IV.A, Theorem IV.1 proof] The proof sketch claims that estimating f² to ε^{2/3} accuracy and then thresholding at 2ε^{2/3} suffices for ε-accurate estimation of |f|. This is incorrect. For example, if |f|=0.5 and ε=0.1, then ε^{2/3}≈0.215, and an allowed error of that size in f² moves the estimated |f| by up to about 0.31, far exceeding ε; near f=0, an estimated square of ε^{2/3} gives an absolute-value estimate of about ε^{1/3}≫ε. To obtain ε accuracy on |f|, the protocol must estimate f² to O(ε²) accuracy, which is consistent with the stated O(ε^{-4}) sample count but is not what the proof says. Because Theorem IV.1 is the basis for the 'easy' side of the Fig. 4 hierarchy and for the tightness claims in Theorems IV.6 and IV.7, the proof as written does not establish the theorem.
  2. [Appendix B, Lemma B.7 and Fig. 5] Lemma B.7 asserts that for a Gaussian γ∈R^m with spread σ, Pr(2σ² ≤ |γ|² ≤ mσ²/0.99) ≥ 0.546 for all m≥8. The proof verifies this numerically on the range 8≤m≤34000 using Fig. 5 and provides an analytic argument only for m≥34000. Since Theorems IV.3, IV.5, B.12, and B.15 use this lemma to guarantee their constant success probability (1/2 + 0.298/3), the bosonic exponential lower bounds are not fully proven for all m≥8 as stated. Please supply an analytic proof or a rigorous interval-arithmetic certificate for the full range, or explicitly present the constant as a numerical conjecture and adjust the affected theorem statements accordingly.
minor comments (4)
  1. [Abstract] The abstract contains grammatical slips: 'Quantum channels can characterized by' should be 'can be characterized by', and 'we bounds tighter lower bounds' should be 'we obtain tighter lower bounds'.
  2. [Lemma III.1, Eq. (19)] The summation index in Eq. (19) is written 'xxx∈{0,...,l}^c, |x|>0', but the vector is xxx; please use a consistent multi-index notation such as |xxx|=Σ_i x_i and state the condition as |xxx|>0.
  3. [Figure 4 caption] The caption says 'self-conjugate channels ... are 1-copy hard, but aren't 2-copy hard'; the intended meaning is 'are not 2-copy hard' (i.e., become easy at two copies), and the term 'hard' should be defined in the caption as exponential in the system size.
  4. [Eq. (6)] The relation \hat D_{d,m}(-q,p) = (\hat D_{d,m}(q,p))^T is stated before the transposition basis is introduced; please add the caveat that the transpose is taken in the Z-diagonal basis to avoid confusion with the later bosonic transposition convention.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the master lemma is proved against an independent discrimination benchmark and the hardness/upper bounds follow from explicit constructions; the Theorem IV.1 accuracy-passing issue is a correctness gap, not circularity.

full rationale

The central lower bounds are not circular. Lemma III.1 is a general statement about c-copy channel-discrimination protocols, proved in Appendix B1 using Radon-Nikodym derivatives, a one-sided Le Cam bound, and χ²-divergence arguments adapted from the external method of Ref. [24]. The reduction from the channel-learning task (Problem II.1) to the many-one discrimination task (Problem III.1) is explicit and one-way: if the learner could solve Problem II.1, it could also solve the revealed discrimination problem, so the lower bound transfers. This is not a definitional equivalence, and the discrimination benchmark is not defined in terms of the learning task. The hardness theorems IV.2-IV.7 instantiate Lemma III.1 with explicitly constructed channel families (Eqs. (32)-(33)), and the relevant operator norms are proved analytically in Lemmas B.8, B.9, B.16, and B.17. The upper bounds are also constructive: Theorem IV.1 uses the external [12,13] Bell-measurement schemes over E⊗E*, and Theorem B.24 uses the external d-copy scheme of [11]. No fitted parameter is later renamed as a prediction. The two self-citations in the paper, Refs. [5] and [43], are contextual (noise benchmarking and GKP error-correcting codes) and are not load-bearing for any claimed bound. No uniqueness theorem from the authors is invoked to forbid alternative constructions. The definitional choices in Problem II.1—estimating absolute values and revealing queries only after all channel copies are consumed—are scope-setting conventions, and the paper explicitly discusses their limitation for bosonic sign and phase estimation; they are not derived conclusions. The proof sketch of Theorem IV.1 contains a genuine accuracy-propagation error: estimating f² to ε^{2/3} accuracy does not generally suffice for ε accuracy in |f| near zero. However, this is a correctness or repairability concern about the stated upper-bound proof, not a circularity: the bound is still attached to an external estimation protocol and is not equivalent to its own input by construction. Therefore no circular step is present.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The ledger shows three hand-picked numerical constants (0.99, 2.42, and the 0.546 success probability), none of which affect asymptotic scaling; six axioms, of which four are standard domain assumptions or standard mathematics; and no invented entities. The two axioms that most limit the strength of the claims are the weakened bosonic task definition (TMSV Choi-state learning, admitted by the authors to be 'weaker' than full characterization, App. C1) and the numerically verified Gaussian concentration bound (Lemma B.7). The paper states its assumptions with unusual honesty, including a no-go result for bosonic transfer-function learning and an example bounding TMSV Choi-state trace distance versus the energy-constrained diamond norm (App. C2).

free parameters (3)
  • Gaussian concentration constants (Lemma B.7) = 0.99 pre-factor; 0.546 probability
    The 0.99 factor in the effective mode dimensions d_in = sqrt(1 + (0.99κ tanh(2r))^2) and d_out = sqrt(1 + (0.99κ')^2) and the 0.546 success-probability constant are chosen by hand in Lemma B.7 and Thm. B.12 so the Gaussian sampling event 2σ² ≤ |γ|² ≤ mσ²/0.99 occurs with constant probability. These constants affect only the constant factors in the bosonic lower bounds, not the asymptotic scaling.
  • Energy-cutoff threshold for bosonic queries = min(κm, κ'm') ≥ 2.42
    Hand-picked in Thms. IV.3, IV.5, and IV.7 and in App. B (Thms. B.12, B.15) to guarantee the Gaussian concentration event of Lemma B.7 has probability ≥ 0.546 and to ensure tanh²(2r) > 2/(κm) in the self-conjugate bosonic construction. Not fitted to data.
  • TMSV squeezing parameter r = cosh(2r) ≥ max(κm, log(2ε^(-2)))
    A task parameter chosen by the learner for the bosonic probe. The bounds depend on r through d_in. The lower bounds require r large enough that the finite-energy Choi state is sufficiently entangled; this is an assumption on the probe, not a fit to data.
assumptions (6)
  • domain assumption Choi-Jamiolkowski isomorphism: learning the Choi state (or TMSV Choi state) is equivalent to the channel-learning task (Eqs. 15-17)
    Invoked in Sec. II to justify that Problem II.1 (estimation of |C_E| entries) characterizes the channel. For bosonic channels the paper itself notes the TMSV characterization is weaker and requires finite r and knowledge of the output state's phase-space decay (Sec. II; App. C1).
  • domain assumption Query set is revealed only after all channel copies are used (Problem II.1)
    This modeling choice makes the lower bounds apply for all queries simultaneously. If queries were known in advance, bosonic sign and phase estimation would be easier, as the paper notes in Sec. II citing [39].
  • ad hoc to paper Lemma B.7: for Gaussian γ in C^m, Pr(2σ² ≤ |γ|² ≤ mσ²/0.99) ≥ 0.546 for all m ≥ 8
    Used to guarantee the constant success probability of the hard bosonic channel instances (Thms. IV.3, IV.5, B.12, B.15). The claim is verified numerically for 8 ≤ m ≤ 34000 via Fig. 5 and analytically for m ≥ 34000 via tail bounds; the finite-range numerical check is not a closed-form proof.
  • standard math Standard measure-theoretic treatment of POVMs with infinite accuracy (Radon-Nikodym, χ² divergence, Le Cam two-point method)
    Used in the proof of the master lemma (App. B1) and in App. A4. The paper explicitly handles unbounded POVM elements by integrating over Borel sets.
  • standard math Heisenberg-Weyl displacement operators form an orthogonal basis (Eqs. 7, 12) and represent all trace-class operators
    Foundation of the transfer matrix and transfer function formalism in Sec. II and Apps. A1-A2. For bosonic modes the orthogonality holds in the tempered-distribution sense (Eq. 12).
  • domain assumption Hard channels E_{(q1,p1),(q2,p2)} and E_{γ1,γ2} are valid CPTP maps implementable by measure-and-prepare circuits (Fig. 3, Eqs. 32-33)
    The lower bounds apply only to these specific channel families. The paper provides explicit constructions via two-outcome POVMs and parity operations, so this is a stated and physically motivated assumption, valid for prime d and for d = 2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Quantum channel learning with limited parallel access." pith.science (2026). https://pith.science/paper/74DGIFX4

@misc{pith2026260805307,
  author       = {Pith},
  title        = {Pith review of: Quantum channel learning with limited parallel access},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/74DGIFX4}},
  note         = {Machine review of arXiv:2608.05307}
}
abstract

Quantum channels can characterized by their action on an orthogonal operator basis, where these operators are related to observable properties of the quantum system. For qudit and multimode bosonic systems, this is encoded respectively in the Heisenberg--Weyl transfer matrix estimated from the Choi state, and in the characteristic-function transfer function estimated from the Choi state generated by probing with a two-mode squeezed vacuum state. We derive sample-complexity bounds for estimating entries of these transfer matrix/function to additive accuracy $\epsilon$ with success probability $\geq1-\delta$, under different resources: access to the complex-conjugate channel $\mathcal{E}^*$ and/or simultaneous access to $c$ copies of the channel. In all settings, the learner uses parallel channel calls with adaptively chosen, ancilla-assisted input states and measurements. Absolute values of transfer-matrix entries can be learned efficiently with simultaneous access to $\mathcal{E}$ and $\mathcal{E}^*$, with tight scaling $\epsilon^{-4}$. Without conjugate access, any $c<d$ copies are insufficient for efficient learning, requiring sample complexity exponential in the number of ($d$-level) qudits $n$ (for prime $d$). Efficiency is recovered at $c=d$, with tight scaling $\epsilon^{-2d}$. For bosonic systems, exponential sample complexity persists for all $c=O(1/\epsilon)$. Although the task is learning a particular state, these bounds carry stronger implications than standard state-learning bounds since the learner controls the inputs and ancillary assistance. This establishes a hierarchy of channel-learning resources: self-complex-conjugate channels require two-copy ancilla-assisted access for efficient learning, while for every square-free $d$, some channels require $d$-copy access. As a corollary, we bounds tighter lower bounds for state learning with limited multi-copy access.

Figures

Figures reproduced from arXiv: 2608.05307 by the authors.

Figure 1
Figure 1. FIG. 1: Representation of a learning protocol that uses [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2: (a) Depiction of [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3: (a) Circuit to implement the two-outcome POVM [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: FIG. 4: Visual representation of learning hierarchy over [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5: Manual verification for lemma [PITH_FULL_IMAGE:figures/full_fig_p035_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 18 canonical work pages

  1. [1]

    author author D. W. \ Leung , author M. A. \ Nielsen , author I. L. \ Chuang ,\ and\ author Y. Yamamoto ,\ title title Approximate quantum error correction can lead to better codes ,\ https://doi.org/10.1103/PhysRevA.56.2567 journal journal Phys. Rev. A \ volume 56 ,\ pages 2567 ( year 1997 ) NoStop

    doi:10.1103/physreva.56.2567

  2. [2]

    author author A. S. \ Fletcher , author P. W. \ Shor ,\ and\ author M. Z. \ Win ,\ title title Optimum quantum error recovery using semidefinite programming ,\ https://doi.org/10.1103/PhysRevA.75.012338 journal journal Phys. Rev. A \ volume 75 ,\ pages 012338 ( year 2007 ) NoStop

    work page doi:10.1103/physreva.75.012338

  3. [3]

    author author D. K. \ Tuckett , author S. D. \ Bartlett ,\ and\ author S. T. \ Flammia ,\ title title Ultrahigh error threshold for surface codes with biased noise ,\ https://doi.org/10.1103/PhysRevLett.120.050505 journal journal Phys. Rev. Lett. \ volume 120 ,\ pages 050505 ( year 2018 ) NoStop

    doi:10.1103/physrevlett.120.050505

  4. [4]

    Bausch , author A

    author author J. Bausch , author A. W. \ Senior , author F. J. H. \ Heras , author T. Edlich , author A. Davies , author M. Newman , author C. Jones , author K. Satzinger , author M. Y. \ Niu , author S. Blackwell , author G. Holland , author D. Kafri , author J. Atalaya , author C. Gidney , author D. Hassabis , author S. Boixo , author H. Neven ,\ and\ a...

    doi:10.1038/s41586-024-08148-8

  5. [5]

    Quantum advantages for syndrome-aware noisy logical observable estimation

    author author K. Tsubouchi , author H. Kwon , author L. Jiang ,\ and\ author N. Yoshioka ,\ https://arxiv.org/abs/2603.05145 title Quantum advantages for syndrome-aware noisy logical observable estimation ( year 2026 ),\ https://arxiv.org/abs/2603.05145 arXiv:2603.05145 [quant-ph] NoStop

    work page Pith review arXiv 2026

  6. [6]

    Carrasco , author A

    author author J. Carrasco , author A. Elben , author C. Kokail , author B. Kraus ,\ and\ author P. Zoller ,\ title title Theoretical and experimental perspectives of quantum verification ,\ https://doi.org/10.1103/PRXQuantum.2.010102 journal journal PRX Quantum \ volume 2 ,\ pages 010102 ( year 2021 ) NoStop

    work page doi:10.1103/prxquantum.2.010102

  7. [7]

    van den Berg , author Z

    author author E. van den Berg , author Z. K. \ Minev , author A. Kandala ,\ and\ author K. Temme ,\ title title Probabilistic error cancellation with sparse pauli--lindblad models on noisy quantum processors ,\ https://doi.org/10.1038/s41567-023-02042-2 journal journal Nature Physics \ volume 19 ,\ pages 1116 ( year 2023 ) NoStop

    doi:10.1038/s41567-023-02042-2

  8. [8]

    Tsubouchi , author T

    author author K. Tsubouchi , author T. Sagawa ,\ and\ author N. Yoshioka ,\ title title Universal cost bound of quantum error mitigation based on quantum estimation theory ,\ https://doi.org/10.1103/PhysRevLett.131.210601 journal journal Phys. Rev. Lett. \ volume 131 ,\ pages 210601 ( year 2023 ) NoStop

    doi:10.1103/physrevlett.131.210601

Show all 60 references
  1. [9]

    Strikis , author D

    author author A. Strikis , author D. Qin , author Y. Chen , author S. C. \ Benjamin ,\ and\ author Y. Li ,\ title title Learning-based quantum error mitigation ,\ https://doi.org/10.1103/PRXQuantum.2.040330 journal journal PRX Quantum \ volume 2 ,\ pages 040330 ( year 2021 ) NoStop

    doi

  2. [10]

    \ Huang , author R

    author author H.-Y. \ Huang , author R. Kueng ,\ and\ author J. Preskill ,\ title title Information-theoretic bounds on quantum advantage in machine learning ,\ https://doi.org/10.1103/PhysRevLett.126.190505 journal journal Phys. Rev. Lett. \ volume 126 ,\ pages 190505 ( year ...

    doi

  3. [11]

    N\"oller , author V

    author author J. N\"oller , author V. T. \ Tran , author M. Gachechiladze ,\ and\ author R. Kueng ,\ https://arxiv.org/abs/2510.08070 title An infinite hierarchy of multi-copy quantum learning tasks ( year 2025 ),\ https://arxiv.org/abs/2510.08070 arXiv:2510.08070 [quant-ph] NoStop

    2025

  4. [12]

    King , author K

    author author R. King , author K. Wan ,\ and\ author J. R. \ McClean ,\ title title Exponential learning advantages with conjugate states and minimal quantum memory ,\ https://doi.org/10.1103/PRXQuantum.5.040301 journal journal PRX Quantum \ volume 5 ,\ pages 040301 ( year 202...

    doi

  5. [13]

    Coroi \ and\ author C

    author author E. Coroi \ and\ author C. Oh ,\ https://arxiv.org/abs/2501.17633 title Exponential advantage in continuous-variable quantum state learning ( year 2025 ),\ https://arxiv.org/abs/2501.17633 arXiv:2501.17633 [quant-ph] NoStop

    2025

  6. [14]

    Chen , author C

    author author S. Chen , author C. Oh , author S. Zhou , author H.-Y. \ Huang ,\ and\ author L. Jiang ,\ title title Tight bounds on pauli channel learning without entanglement ,\ https://doi.org/10.1103/PhysRevLett.132.180805 journal journal Phys. Rev. Lett. \ volume 132 ,\ pa...

    doi

  7. [15]

    Seif , author S

    author author A. Seif , author S. Chen , author S. Majumder , author H. Liao , author D. S. \ Wang , author M. Malekakhlagh , author A. Javadi-Abhari , author L. Jiang ,\ and\ author Z. K. \ Minev ,\ https://arxiv.org/abs/2408.03376 title Entanglement-enhanced learning of quan...

    2024 arXiv

  8. [16]

    Chen , author S

    author author S. Chen , author S. Zhou , author A. Seif ,\ and\ author L. Jiang ,\ title title Quantum advantages for pauli channel estimation ,\ https://doi.org/10.1103/PhysRevA.105.032435 journal journal Phys. Rev. A \ volume 105 ,\ pages 032435 ( year 2022 ) NoStop

    doi

  9. [17]

    Oh , author S

    author author C. Oh , author S. Chen , author Y. Wong , author S. Zhou , author H.-Y. \ Huang , author J. A. H. \ Nielsen , author Z.-H. \ Liu , author J. S. \ Neergaard-Nielsen , author U. L. \ Andersen , author L. Jiang ,\ and\ author J. Preskill ,\ title title Entanglement-...

    doi

  10. [18]

    \ Liu , author R

    author author Z.-H. \ Liu , author R. Brunel , author E. E. B. \ Østergaard , author O. Cordero , author S. Chen , author Y. Wong , author J. A. H. \ Nielsen , author A. B. \ Bregnsbo , author S. Zhou , author H.-Y. \ Huang , author C. Oh , author L. Jiang , author J. Preskill...

    doi

  11. [19]

    Kunjummen , author M

    author author J. Kunjummen , author M. C. \ Tran , author D. Carney ,\ and\ author J. M. \ Taylor ,\ title title Shadow process tomography of quantum channels ,\ https://doi.org/10.1103/PhysRevA.107.042403 journal journal Phys. Rev. A \ volume 107 ,\ pages 042403 ( year 2023 ) NoStop

    doi

  12. [20]

    Li , author C

    author author Z. Li , author C. Yi , author Y. Zhou ,\ and\ author H. Zhu ,\ title title Nearly query-optimal classical shadow estimation of unitary channels ,\ https://doi.org/10.1103/wccm-zys6 journal journal PRX Quantum \ volume 6 ,\ pages 030366 ( year 2025 ) NoStop

    doi

  13. [21]

    Levy , author D

    author author R. Levy , author D. Luo ,\ and\ author B. K. \ Clark ,\ title title Classical shadows for quantum process tomography on near-term quantum computers ,\ https://doi.org/10.1103/PhysRevResearch.6.013029 journal journal Phys. Rev. Res. \ volume 6 ,\ pages 013029 ( ye...

    doi

  14. [22]

    He , author Z

    author author E. He , author Z. Li , author N. Scully , author S. Zhou ,\ and\ author Y. Yang ,\ https://arxiv.org/abs/2606.13638 title Optimal classical shadow estimation of unitary channels at heisenberg limit ( year 2026 ),\ https://arxiv.org/abs/2606.13638 arXiv:2606.13638...

    2026 arXiv

  15. [23]

    author author M. C. \ Caro ,\ title title Learning quantum processes and hamiltonians via the pauli transfer matrix ,\ journal journal ACM Transactions on Quantum Computing \ volume 5 ,\ https://doi.org/10.1145/3670418 10.1145/3670418 ( year 2024 ) NoStop

    doi

  16. [24]

    Chen , author W

    author author S. Chen , author W. Gong ,\ and\ author Q. Ye ,\ https://arxiv.org/abs/2404.19105 title Optimal tradeoffs for estimating pauli observables ( year 2024 b ),\ https://arxiv.org/abs/2404.19105 arXiv:2404.19105 [quant-ph] NoStop

    2024 arXiv

  17. [25]

    author author A. A. \ Mele \ and\ author L. Bittel ,\ https://arxiv.org/abs/2512.10214 title Optimal learning of quantum channels in diamond distance ( year 2026 ),\ https://arxiv.org/abs/2512.10214 arXiv:2512.10214 [quant-ph] NoStop

    2026

  18. [26]

    Walter \ and\ author F

    author author M. Walter \ and\ author F. Witteveen ,\ https://arxiv.org/abs/2512.15690 title A random purification channel for arbitrary symmetries with applications to fermions and bosons ( year 2025 ),\ https://arxiv.org/abs/2512.15690 arXiv:2512.15690 [quant-ph] NoStop

    2025

  19. [27]

    Tang , author J

    author author E. Tang , author J. Wright ,\ and\ author M. Zhandry ,\ https://arxiv.org/abs/2510.07622 title Conjugate queries can help ( year 2026 ),\ https://arxiv.org/abs/2510.07622 arXiv:2510.07622 [quant-ph] NoStop

    2026 arXiv

  20. [28]

    Girardi , author F

    author author F. Girardi , author F. A. \ Mele ,\ and\ author L. Lami ,\ https://arxiv.org/abs/2511.23451 title Random purification channel made simple ( year 2026 ),\ https://arxiv.org/abs/2511.23451 arXiv:2511.23451 [quant-ph] NoStop

    2026 arXiv

  21. [29]

    author author F. A. \ Mele , author A. A. \ Mele , author L. Bittel , author J. Eisert , author V. Giovannetti , author L. Lami , author L. Leone ,\ and\ author S. F. E. \ Oliviero ,\ title title Learning quantum states of continuous variable systems ,\ https://doi.org/10.1038...

  22. [30]

    author author B. C. \ Hall ,\ @noop title Quantum theory for mathematicians \ ( publisher Springer ,\ year 2013 ) NoStop

    2013

  23. [31]

    \ Huang , author S

    author author H.-Y. \ Huang , author S. Chen ,\ and\ author J. Preskill ,\ title title Learning to predict arbitrary quantum processes ,\ https://doi.org/10.1103/PRXQuantum.4.040337 journal journal PRX Quantum \ volume 4 ,\ pages 040337 ( year 2023 ) NoStop

    doi

  24. [32]

    Sarkar \ and\ author T

    author author R. Sarkar \ and\ author T. J. \ Yoder ,\ title title The qudit P auli group: non-commuting pairs, non-commuting sets, and structure theorems ,\ https://doi.org/10.22331/q-2024-04-04-1307 journal journal Quantum \ volume 8 ,\ pages 1307 ( year 2024 ) NoStop

    2024 doi

  25. [33]

    Becker , author N

    author author S. Becker , author N. Datta , author L. Lami ,\ and\ author C. Rouze ,\ title title Classical shadow tomography for continuous variables quantum systems ,\ https://doi.org/10.1109/TIT.2024.3357972 journal journal IEEE Transactions on Information Theory \ volume 7...

    2024

  26. [34]

    An example is the Dirac delta function

    note A tempered distribution is a continuous linear functional on the space of Schwartz functions. An example is the Dirac delta function. Stop

  27. [35]

    More information on this can be found in moretti2017spectral NoStop

    note This is a C^* algebra W defined for symplectic space (X, ) such that the generators of this algebra satisfy for all f,g X we have W(f)^ = W(-f) and W(f)W(g) = W(f+g)e^ i (g,f)/2 . More information on this can be found in moretti2017spectral NoStop

  28. [36]

    Serafini ,\ @noop title Quantum continuous variables: a primer of theoretical methods \ ( publisher CRC press ,\ year 2023 ) NoStop

    author author A. Serafini ,\ @noop title Quantum continuous variables: a primer of theoretical methods \ ( publisher CRC press ,\ year 2023 ) NoStop

    2023

  29. [37]

    King , author D

    author author R. King , author D. Gosset , author R. Kothari ,\ and\ author R. Babbush ,\ title Triply efficient shadow tomography ,\ in\ https://doi.org/10.1137/1.9781611978322.27 booktitle Proceedings of the 2025 Annual ACM-SIAM Symposium on Discrete Algorithms (SODA) ,\ pp....

    2025 doi

  30. [38]

    author author V. T. \ Tran \ and\ author R. Kueng ,\ title title One, two, three: One empirical evaluation of a two-copy shadow tomography scheme with triple efficiency ,\ in\ https://doi.org/10.1109/QCE65121.2025.00177 booktitle 2025 IEEE International Conference on Quantum C...

    2025

  31. [39]

    \ Wu , author Y

    author author Y.-D. \ Wu , author Y. Zhu , author G. Chiribella ,\ and\ author N. Liu ,\ title title Efficient learning of continuous-variable quantum states ,\ https://doi.org/10.1103/PhysRevResearch.6.033280 journal journal Phys. Rev. Res. \ volume 6 ,\ pages 033280 ( year 2...

    doi

  32. [40]

    Zhu , author Z

    author author C. Zhu , author Z. Tang , author G. Zhen , author Y. Li , author G. Bai ,\ and\ author X. Wang ,\ https://arxiv.org/abs/2602.05828 title Simulation of adjoints and petz recovery maps for unknown quantum channels ( year 2026 ),\ https://arxiv.org/abs/2602.05828 ar...

    2026

  33. [41]

    Busch ,\ title title Unsharp reality and joint measurements for spin observables ,\ https://doi.org/10.1103/PhysRevD.33.2253 journal journal Phys

    author author P. Busch ,\ title title Unsharp reality and joint measurements for spin observables ,\ https://doi.org/10.1103/PhysRevD.33.2253 journal journal Phys. Rev. D \ volume 33 ,\ pages 2253 ( year 1986 ) NoStop

    1986 doi

  34. [42]

    Conrad , author J

    author author J. Conrad , author J. Eisert ,\ and\ author F. Arzani ,\ title title Gottesman- Kitaev - Preskill codes: A lattice perspective ,\ https://doi.org/10.22331/q-2022-02-10-648 journal journal Quantum \ volume 6 ,\ pages 648 ( year 2022 ) ,\ note arXiv:2109.14645 [qua...

    2022 arXiv

  35. [43]

    Subramanian , author G

    author author M. Subramanian , author G. Zheng ,\ and\ author L. Jiang ,\ title title Achievable rates for concatenated square gottesman-kitaev-preskill codes ,\ https://doi.org/10.1103/56vj-z7h1 journal journal PRX Quantum \ volume 6 ,\ pages 040342 ( year 2025 ) NoStop

    doi

  36. [44]

    author author S. P. \ Jain , author J. T. \ Iosue , author A. Barg ,\ and\ author V. V. \ Albert ,\ title title Quantum spherical codes ,\ https://doi.org/10.1038/s41567-024-02496-y journal journal Nature Physics \ volume 20 ,\ pages 1300 ( year 2024 ) NoStop

    doi

  37. [45]

    Aharonov , author A

    author author D. Aharonov , author A. Kitaev ,\ and\ author N. Nisan ,\ https://arxiv.org/abs/quant-ph/9806029 title Quantum circuits with mixed states ( year 1998 ),\ https://arxiv.org/abs/quant-ph/9806029 arXiv:quant-ph/9806029 [quant-ph] NoStop

    1998 arXiv

  38. [46]

    Oufkir \ and\ author F

    author author A. Oufkir \ and\ author F. Girardi ,\ https://arxiv.org/abs/2601.04180 title Improved lower bounds for learning quantum channels in diamond distance ( year 2026 ),\ https://arxiv.org/abs/2601.04180 arXiv:2601.04180 [quant-ph] NoStop

    2026

  39. [47]

    Haah , author R

    author author J. Haah , author R. Kothari , author R. O’Donnell ,\ and\ author E. Tang ,\ title title Query-optimal estimation of unitary channels in diamond distance ,\ in\ https://doi.org/10.1109/FOCS57990.2023.00028 booktitle 2023 IEEE 64th Annual Symposium on Foundations o...

    2023

  40. [48]

    Fanizza , author V

    author author M. Fanizza , author V. Iyer , author J. Lee , author A. A. \ Mele ,\ and\ author F. A. \ Mele ,\ https://arxiv.org/abs/2510.05531 title Efficient learning of bosonic gaussian unitaries ( year 2025 ),\ https://arxiv.org/abs/2510.05531 arXiv:2510.05531 [quant-ph] NoStop

    2025 arXiv

  41. [49]

    author author M. E. \ Shirokov ,\ title title On the energy-constrained diamond norm and its application in quantum information theory ,\ https://doi.org/10.1134/S0032946018010027 journal journal Problems of Information Transmission \ volume 54 ,\ pages 20 ( year 2018 ) NoStop

    doi

  42. [50]

    author author A. Winter ,\ https://arxiv.org/abs/1712.10267 title Energy-constrained diamond norm with applications to the uniform continuity of continuous variable channel capacities ( year 2017 ),\ https://arxiv.org/abs/1712.10267 arXiv:1712.10267 [quant-ph] NoStop

    2017 arXiv

  43. [51]

    Bavaresco , author M

    author author J. Bavaresco , author M. Murao ,\ and\ author M. T. \ Quintino ,\ title title Strict hierarchy between parallel, sequential, and indefinite-causal-order strategies for channel discrimination ,\ https://doi.org/10.1103/PhysRevLett.127.200504 journal journal Phys. ...

    doi

  44. [52]

    Chen \ and\ author W

    author author S. Chen \ and\ author W. Gong ,\ title title Efficient pauli channel estimation with logarithmic quantum memory ,\ https://doi.org/10.1103/PRXQuantum.6.020323 journal journal PRX Quantum \ volume 6 ,\ pages 020323 ( year 2025 ) NoStop

    doi

  45. [53]

    Moretti ,\ @noop title Spectral theory and quantum mechanics \ ( publisher Springer ,\ year 2017 ) NoStop

    author author V. Moretti ,\ @noop title Spectral theory and quantum mechanics \ ( publisher Springer ,\ year 2017 ) NoStop

    2017

  46. [54]

    Fl\"uhmann \ and\ author J

    author author C. Fl\"uhmann \ and\ author J. P. \ Home ,\ title title Direct characteristic-function tomography of quantum states of the trapped-ion motional oscillator ,\ https://doi.org/10.1103/PhysRevLett.125.043602 journal journal Phys. Rev. Lett. \ volume 125 ,\ pages 043...

    doi

  47. [55]

    Asadian , author P

    author author A. Asadian , author P. Erker , author M. Huber ,\ and\ author C. Kl\"ockl ,\ title title Heisenberg-weyl observables: Bloch vectors in phase space ,\ https://doi.org/10.1103/PhysRevA.94.010301 journal journal Phys. Rev. A \ volume 94 ,\ pages 010301 ( year 2016 ) NoStop

    doi

  48. [56]

    author author V. P. \ Gupta , author P. Mandayam ,\ and\ author V. Sunder ,\ title title The functional analysis of quantum information theory ,\ @noop journal journal Lecture Notes in Physics \ volume 902 ( year 2015 ) NoStop

    2015

  49. [57]

    author author V. I. \ Bogachev \ and\ author M. A. S. \ Ruas ,\ @noop title Measure theory ,\ Vol. volume 1 \ ( publisher Springer ,\ year 2007 ) NoStop

    2007

  50. [58]

    author author M. Ghosh ,\ title title Exponential tail bounds for chisquared random variables ,\ @noop journal journal Journal of Statistical Theory and Practice \ volume 15 ,\ pages 35 ( year 2021 ) NoStop

    2021

  51. [59]

    author author M. H. \ Shaw , author A. C. \ Doherty ,\ and\ author A. L. \ Grimsmo ,\ title title Stabilizer subsystem decompositions for single- and multimode gottesman-kitaev-preskill codes ,\ https://doi.org/10.1103/PRXQuantum.5.010331 journal journal PRX Quantum \ volume 5...

    doi

  52. [60]

    Bittel , author F

    author author L. Bittel , author F. A. \ Mele , author A. A. \ Mele , author S. Tirone ,\ and\ author L. Lami ,\ title title Optimal estimates of trace distance between bosonic G aussian states and applications to learning ,\ https://doi.org/10.22331/q-2025-06-12-1769 journal ...

    doi

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.