Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Bayesian frequency estimation at the fundamental quantum limit

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read By projecting onto quantum-whitened states, one measurement attains the fundamental Bayesian quantum limit for frequency estimation in Toeplitz signal families.

desk verdict A clean closed-form Bayesian limit for Toeplitz families, but the practical frequency-estimation advantage is drawn from finite-prior numerics where the proposed measurement is not the proven optimum. read the letter →

arxiv 2507.02811 v2 pith:IQ7M4VM2 submitted 2025-07-03 quant-ph gr-qc

James W. Gardner , Tuvia Gefen , Ethan Payne , Su Direkci , Sander M. Vermeulen , Simon A. Haine , Joseph J. Hope , Lee McCuller
show 1 more author
Yanbei Chen
This is my paper · ORCID
classification quant-phgr-qc
keywords BayesianquantumestimationfrequencywhiteningToeplitzfamilycoherentstatescovariantmeasurementSNRthresholdmetrology
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

This paper establishes the fundamental Bayesian quantum limit for estimating a parameter encoded in any family of pure states whose overlaps depend only on the parameter difference, a Toeplitz family. It proves that, with an exactly flat normalized prior, the smallest possible mean-square error is $V=\frac{1}{4}\int_S \frac{g'(k)^2}{g(k)}\,dk$, where $g(k)$ is the spectral measure of the overlap function, and that this error is attained by quantum whitening: projecting the state onto the Fourier-conjugate basis built from the Gram matrix's spectrum. Applied to coherent-state frequency estimation of a sinusoid, the Gram matrix becomes Toeplitz in the many-cycles limit, so whitening is optimal and numerically beats time-domain quadrature measurement below the classical signal-to-noise threshold. This matters because weak-signal searches for gravitational-wave remnants and axion dark matter are limited by exactly this frequency-localisation threshold.

What carries the argument

The load-bearing object is the quantum-whitening measurement. Given a Toeplitz family with overlap $G(\theta-\theta')=\langle\psi_\theta|\psi_{\theta'}\rangle$, one Fourier-transforms and normalises by $\sqrt{g(k)}$ to build the orthogonal Fourier states $|k\rangle$, then forms the covariant states $|\theta\rangle=\frac{1}{\sqrt{2\pi}}\int_S dk\, e^{-ik\theta}|k\rangle$. Measuring $\hat{W}=\int d\theta\, \theta\,|\theta\rangle\langle\theta|$ projects onto these whitened states. For a flat prior this operator satisfies the Bayesian Lyapunov equation, so it is the Bayesian symmetric logarithmic derivative; the calculation reduces the MBMSE to a Fisher-information-like functional of the spectral measure $g(k)$.

What would settle it

Compute the numerical MBMSE from the Bayesian symmetric logarithmic derivative for a finite flat prior with $\Delta\omega=2\pi\times0.9$ Hz, $T=10$ s, and SNR $\approx4.5$, using the exact non-Toeplitz Gram matrix of Eq. 70, and compare it with the Bayesian MSE of quantum whitening; if whitening's MSE exceeds the MBMSE by more than the numerical tolerance, its optimality does not extend to finite priors. Alternatively, exhibit any Toeplitz family with smooth spectral measure $g$ and a measurement whose flat-prior BMSE is smaller than $\frac{1}{4}\int_S g'(k)^2/g(k)\,dk$, which would disprove the theorem.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that for any Toeplitz family of pure states $\{|\psi_\theta\rangle\}$ with a normalized flat prior, the minimum Bayesian mean-square error is $V=\frac{1}{4}\int_S \frac{g'(k)^2}{g(k)}\,dk$, and the optimal measurement is the covariant one: project onto the quantum-whitened states $|\theta\rangle=\frac{1}{\sqrt{2\pi}}\int_S dk\, e^{-ik\theta}|k\rangle$. The argument diagonalises the prior-averaged state in the Fourier basis of the family and shows that the whitening operator equals the Bayesian symmetric logarithmic derivative, so no other measurement can improve on it. For coherent-state frequency estimation of a sinusoid, the many-cycles limit makes the Gram matrix Toeplitz, so whitening becomes the optimal measurement; numerically it beats time-domain quadrature (and other practical schemes) below the SNR threshold, while the threshold itself persists at the quantum level. The paper also shows that for an infinitely wide flat prior the frequency-estimation limit diverges, so the practical advantage is established numerically for finite-width priors.

Load-bearing premise

The proof of optimality assumes an exactly flat prior spread over all real frequencies, which is a formal idealisation, and the practical advantage for sinusoids also relies on the many-cycles approximation that makes the Gram matrix Toeplitz; for finite priors the paper supplies numerics rather than proof.

Editorial extensions

If this is right

  • For any Toeplitz family under a flat normalized prior, quantum whitening attains the MBMSE $V=\frac{1}{4}\int_S g'(k)^2/g(k)\,dk$, so no measurement can beat it.
  • In coherent-state frequency estimation in the many-cycles limit, quantum whitening is optimal and beats time-domain quadrature and filter-based photon-counting strategies below the SNR threshold.
  • The classical SNR threshold effect persists at the fundamental quantum limit: below threshold the MBMSE sits far above the quantum Cramér-Rao bound.
  • The infinite-flat-prior limit diverges for frequency estimation, so for wide-but-finite priors the achievable gain is quantified by the ratio of posterior to prior variance, which the numerics show can be substantial.
  • The result generalises to any unitary encoding with a Toeplitz overlap structure, giving a parameter-agnostic sensing strategy when no prior information is available.

Reading between the lines

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

  • One implicit consequence is that the difficulty of no-prior parameter estimation is controlled by the spectral measure $g(k)$: the limit $V$ is a Fisher-information-like functional of $g$, so encodings with smoother, lighter-tailed spectral measures should be easier to sense, a design principle the paper does not state explicitly.
  • A testable extension is to run the same Bayesian optimisation for non-Toeplitz families with finite priors and compare whitening to the numerical MBMSE; the paper reports small residual differences, and tightening this comparison could reveal whether a modified covariant measurement closes the gap.
  • The divergence of $V$ in the infinite-prior limit suggests practical sensitivity statements should report posterior-variance reduction relative to prior width; this ratio, rather than an absolute MSE, could become the standard figure for single-event frequency searches.
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 / 4 minor

Summary. The paper derives a Bayesian quantum limit for estimating a continuous parameter encoded in a Toeplitz family of pure states under a formally normalized flat prior. The central result is Eq. (34), V = (1/4) ∫_S [g'(k)^2/g(k)] dk, and the claimed optimal measurement is the "quantum whitening" POVM that projects onto the states |θ⟩ of Eq. (24). The authors apply this framework to coherent-state frequency estimation, arguing that in the many-cycles limit the Gram matrix becomes Toeplitz (Eq. 72), that quantum whitening is a covariant measurement, and that numerically the Bayesian-optimal measurement beats time-domain quadrature below the SNR threshold (Fig. 3a). The paper also treats single-particle states, amplitude and phase estimation, and provides a numerical SVD method for the exact non-Toeplitz Gram matrix.

Significance. The formal machinery is a genuine contribution: for Toeplitz families the paper gives a closed-form MBMSE and identifies a natural covariant POVM, with the proof of Eq. (34) in Appendix B carried out carefully under the stated flat-prior assumptions. The numerical SVD approach for the exact non-Toeplitz Gram matrix is useful and the code/data are advertised as openly available. However, the headline practical claim — that quantum whitening beats quadrature at finite SNR — is not directly demonstrated. The analytic optimality proof applies only to the infinite-flat-prior limit, where the frequency-estimation MBMSE diverges, and Fig. 3a plots the MBMSE of a different, BSLD-optimal POVM rather than the BMSE of quantum whitening. The paper is therefore more convincing as a derivation of the fundamental Bayesian limit plus a candidate measurement than as a demonstrated finite-SNR protocol.

major comments (3)
  1. [Sec. VC3 and Fig. 3a] The abstract and Sec. VI claim that quantum whitening beats quadrature below the SNR threshold, but the finite-prior calculation in Fig. 3a reports the MBMSE achieved by the BSLD-optimal POVM for the exact non-Toeplitz family, which Sec. VC3 explicitly states is not quantum whitening because the wide-prior limit has not been taken. No curve computes the BMSE of the quantum-whitening POVM at finite SNR, so the central practical claim is not verified by the data shown; please add such a curve, or a rigorous bound showing that the whitening BMSE lies below the quadrature curve in the relevant SNR range.
  2. [Sec. IIC1 and Sec. VC1] The analytic optimality proof requires the exactly flat normalized prior on R, expressed informally as ε = 1/[2πδ(0)]. For coherent frequency estimation the spectral measure g(k) has a delta contribution at k = 0 and discontinuities at k = ±T, so Eq. (34) diverges, as the text acknowledges near Eq. (82) and in Sec. VC2. Consequently, the only regime in which quantum whitening is proven optimal for frequency estimation has infinite MSE; the claimed practical advantage rests on the finite-prior numerical regime, where optimality of quantum whitening is explicitly deferred. Please either provide an analytic finite-prior treatment or reformulate the claim as a wide-prior relative-gain result with the divergence stated up front.
  3. [Sec. VC2] The numerical comparison between the whitening posterior variance and the MBMSE is left with three possible explanations for the residual difference, including that quantum whitening is close to but not fully optimal for a wide but finite prior. Since the finite-prior case is exactly where the practical claim lives, the manuscript needs to resolve this residual: for example, compute the exact BMSE of quantum whitening using the true posterior rather than the likelihood, and show numerically that the finite-prior MBMSE converges to the whitening BMSE as the prior width increases. Without this, the gap between the derived optimality result and the finite-SNR advantage claim remains open.
minor comments (4)
  1. [Fig. 3a] The caption and legend mix "BMSE" and "MBMSE" without stating that the MBMSE curve is obtained by numerical BSLD diagonalization of the exact non-Toeplitz Gram matrix, and without noting that no quantum-whitening curve is present; please make this explicit so readers do not conflate the two.
  2. [Sec. IIC1] The statement that "δ(0)/δ(0) = 1" is informal, and Appendix D5 validates the limiting procedure only for the displacement example; please state explicitly in the main text that no analogous finite-prior limit is proved for the frequency case, since this is where the conjecture is introduced.
  3. [Eq. (72)] The many-cycles condition T ≫ 1/min(ω, ω′) should be stated more carefully, since for frequencies that are very close to each other the relevant separation is controlled by the product (ω−ω′)T rather than by the absolute frequency scale; the current wording may mislead readers about the validity of the Toeplitz approximation.
  4. [Ref. [40]] The data/code URL is a git.ligo.org link; please verify that it is publicly accessible without institutional credentials, since the availability statement is part of the reproducibility claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central MBMSE and quantum-whitening optimality results are derived from the Toeplitz/BSLD formalism with no fitted inputs and no load-bearing self-citations.

full rationale

The paper's central derivation is self-contained. Starting from a Toeplitz family, it constructs Fourier states through the spectral measure g(k), proves that any Toeplitz family is generated by an additive unitary, and then computes the BSLD and the MBMSE. The optimality of quantum whitening is shown directly by verifying that the whitening operator W has the same matrix coefficients as the BSLD (Sec. IIC4, Eqs. 42-44), in addition to the standard covariant-measurement argument. No parameter is fitted and then renamed as a prediction; Eq. 34 is obtained by explicit posterior-variance calculation under the stated flat normalized prior. The cited works by the same authors (Refs. [1], [21], [28]) are used for context, contrast, or head-to-head comparison, not as premises of the derivation. The manuscript itself explicitly flags the finite-prior limitation: Sec. VC3 states that the optimal measurement achieving the finite-prior MBMSE 'is not quantum whitening, because we have not taken the limit of wider and wider priors,' and Sec. VC2 lists non-optimality as a possible explanation for the residual difference. That is an honest gap between the abstract's practical claim and the computed quantity, not a circular step in the derivation. The formal epsilon prior is defined and handled with stated care, and the divergences of Eq. 34 for frequency estimation are acknowledged and analyzed. Therefore no circularity is present.

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

No fitted parameters appear anywhere in the derivation; A, T, φ, and the prior width are stated inputs. No new physical entities are introduced: quantum-whitened states are mathematical superpositions inside the existing Hilbert space, not new particles, forces, or dimensions. The main burden is the formal flat-prior construction and the many-cycles approximations needed to make the motivating problem Toeplitz.

assumptions (6)
  • standard math The Bayesian quantum estimation framework: MBMSE = Δ²θ(0) − Tr[ρ̄ L²] with the BSLD solving the Lyapunov equation.
    Invoked in Sec. IA as the starting point; this is a standard result cited from Ref. [13].
  • standard math For a flat prior and a covariant cost function, the optimal measurement is covariant with respect to the symmetry group.
    Used in Sec. IIA and Appendix A2 to identify the quantum-whitening POVM as optimal; this is the standard group-covariant estimation theorem, Refs. [36-38].
  • ad hoc to paper The prior is exactly flat and normalized on R, π(θ)=ε with ε=1/[2πδ(0)], and delta-function identities are used freely.
    Needed to make the mixed state diagonal in the Fourier basis and to make the posterior equal to the likelihood in Sec. IIC. Finite or non-uniform priors are not solved analytically, and the paper conjectures the wide-prior limit.
  • domain assumption The coherent-state Gram matrix for frequency estimation is approximated by the many-cycles Toeplitz form exp[-A²T(1-sinc(ωT))/4].
    Used in Sec. IVC, Eq. 72; requires T≫1/min(ω,ω'), and the exact non-Toeplitz case is only studied numerically in Sec. VC3.
  • domain assumption The signal amplitude A and phase φ are known, and the signal is windowed on (0,T) with the discretization choices of Table III.
    Defines the canonical problem in Sec. IVC and the numerical setup in Appendix K; the authors note that marginalizing over amplitude and phase changes the comparison.
  • domain assumption The spectral measure g(k) is smooth enough that the integration by parts in Appendix B1 is valid.
    The derivation of Eq. 34 requires differentiability of g and vanishing boundary terms; this fails for the sinc symbol, producing the divergence that is central to the frequency-estimation discussion.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Bayesian frequency estimation at the fundamental quantum limit." pith.science (2026). https://pith.science/paper/IQ7M4VM2

@misc{pith2026250702811,
  author       = {Pith},
  title        = {Pith review of: Bayesian frequency estimation at the fundamental quantum limit},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IQ7M4VM2}},
  note         = {Machine review of arXiv:2507.02811}
}
read the original abstract

Searching for a weak signal at an unknown frequency is a canonical task in experiments probing fundamental physics such as gravitational-wave observatories and ultra-light dark matter haloscopes. These state-of-the-art sensors are limited by quantum noise arising from the fundamental uncertainty about the state of the device. Classically, frequency estimation suffers from a threshold effect in the signal-to-noise ratio such that weak signals are extremely hard to localise in frequency. We show that this phenomenon persists at the fundamental quantum limit but that the classical approach, a quadrature measurement, can nevertheless be beaten by a coherent protocol of projecting onto the "quantum whitened" possible quantum states. Quantum whitening is a covariant measurement, and we examine it analytically in the wide-prior limit and numerically for finite-width priors. Beyond accelerating searches for unknown frequencies, quantum whitening may be used generally to sense the parameter of a unitary encoding given no prior information about the parameter.

Figures

Figures reproduced from arXiv: 2507.02811 by the authors.

Figure 1
Figure 1. FIG. 1. A sinusoidal signal [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Exact and approximate (truncated at 7th order) val [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. (a) BMSE in amplitude units for different measure [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Extending the dynamic range in quantum frequency estimation with sequential weak measurements

    quant-ph 2025-09 conditional novelty 6.0 of 10

    Sequential weak measurements followed by a final projective measurement extend the dynamic range of coherent-spin-state frequency estimation and asymptotically saturate the noiseless quantum Fisher information bound.

Reference graph

Works this paper leans on

64 extracted references · 54 canonical work pages · cited by 1 Pith paper

  1. [1]

    Then,g(k) =δ(k)and the MBMSE in Eq

    Constant case Suppose thatG(θ)≡1such that the state is indepen- dent of the parameter and the posterior equals the prior since no information is gained from any measurement. Then,g(k) =δ(k)and the MBMSE in Eq. 34 diverges as it equals the prior variance. Similarly, the QFI in Eq. 36 vanishes such that the QCRB diverges. This is the wide limitς→0of the Gau...

    work page

  2. [2]

    Suppose that G(θ) =δ(θ)/δ(0)such that the states are all orthogo- nal, e.g

    Singular case Let us now consider the opposite case. Suppose that G(θ) =δ(θ)/δ(0)such that the states are all orthogo- nal, e.g. the position eigenstates, theng(k)≡1/[2πδ(0)] which is the uninformative, uniform probability distri- bution onRsimilarly to the prior. The MBMSE in Eq. 34 vanishes as the parameter can be exactly deter- mined from a single meas...

    work page

  3. [3]

    Each term here corresponds to one additional particle: the first term to zero particles, the second to one particle, the third to two particles etc

    Low-SNR expansion These divergences can be explained analytically in the low-SNR limitA 2T≪1since then the symbol is G(ω)≈e − 1 4 A2T 1 + A2T 4 sinc(ωT) + A4T 2 32 sinc2(ωT) (81) where we drop terms of orderO(A6T 3/384). Each term here corresponds to one additional particle: the first term to zero particles, the second to one particle, the third to two pa...

    work page

  4. [4]

    the minimum average posterior variance, and the prior variance for a sequence of wider and wider priors

    Limiting behaviour for wider and wider priors Similarly to the discussion in the sinc case above, this means that the BMSE is a poor figure-of-merit for fre- quency estimation with the uninformative, uniform prior onRand that we want to instead study the ratio be- tween the MBMSE, i.e. the minimum average posterior variance, and the prior variance for a s...

    work page

  5. [5]

    quantum Fourier transform

    Non-Toeplitz case Let us now examine what happens if we do not as- sume that the Gram matrix is Toeplitz by not taking the many-cycles limit. Using the numerical approach in Appendix K, we can study our motivating example of fre- quency estimation of the sinusoidal signal in Eq. 67 with known amplitude and phase and the exact non-Toeplitz Gram matrix in E...

    work page

  6. [6]

    Non-covariant figure-of-merit To show that quantum whitening is not optimal, we need to show thatˆWdoes not solve the Lyapunov equa- tion for the BSLDˆLin Eq. 9. The BMSE in Eq. 5 is not covariant in this case since the MSE is not periodic. This means that although quantum whitening is covariant it need not be optimal here, unlike thek∈Rcase, since the co...

    work page

  7. [7]

    derivative

    Similarly to Eq. 38, the mixed state is thus diagonal in the Fourier basis (the eigenbasis of the Hamiltonian) since ˆ¯ρ= X k∈S gk|φk⟩⟨φk|.(A12) And, instead of Eq. 39, the Bayesian derivative is now ˆ¯ρ′ =i X k,l∈S;k̸=l (−1)k−l k−l √gkgl|φk⟩⟨φl|.(A13) Instead of Eq. 40 or Eq. 42, the BSLD in Eq. 11 fork̸=l is thus ⟨φk| ˆL|φl⟩=i (−1)k−l k−l 2√gkgl gk +g l...

    work page

  8. [8]

    The periodic symmetry of a circu- lant family of states{|ψθ⟩}θ∈(−π,π] means that we should switch from using the MSE in Eq

    Covariant figure-of-merit The issue above is not with the quantum whitening measurement, which is covariant, but with the cost func- tion, which is not. The periodic symmetry of a circu- lant family of states{|ψθ⟩}θ∈(−π,π] means that we should switch from using the MSE in Eq. 4 to some periodic cost function. In the periodic case, the additive unitary is ...

    work page

Show all 64 references
  1. [9]

    The cost function,C(θ1, θ2), is group invariant: C(θ 1, θ2) =C(θ 1 +xmod (−π, π], θ2 +xmod (−π, π]). (A17)

  2. [10]

    An estimation problem that satisfies these conditions is covariant

    The prior distribution is group invariant,π(θ+ xmod (−π, π]) =π(θ). An estimation problem that satisfies these conditions is covariant. It is then well-established that the optimal 19 measurement is also covariant with respect to this group actionevenifitisanon-projectivePOVM[...

  3. [11]

    34 holds for the MBMSE given the uninformative, uniform prior

    BMSE of quantum whitening Here, we show that Eq. 34 holds for the MBMSE given the uninformative, uniform prior. We also show the asso- ciated results in Sec. IIC2. The covariant measurement is optimal such that we need to calculate the BMSE of quantum whitening. First, the lik...

  4. [12]

    Optimality of quantum whitening Here, we provide further details for the indepen- dent proof of the optimality of quantum whitening in Sec. IIC4. The mixed state ˆ¯ρand its Bayesian deriva- tive ˆ¯ρ′ in Eqs. 38 and 39 come directly from applying the result for a general prior ...

  5. [13]

    7 becomes [13] ∆2µ=σ 2 −σ 4I ˆϱ Q(µ)(D1) whereI ˆϱ Q(µ)is the QFI with respect toµof the displaced mixedstateˆϱ=e iµˆxˆ¯ρe−iµˆx

    MBMSE In the case of a Gaussian prior,π(µ) = e−µ2/(2σ2)/(σ √ 2π)with varianceσ 2, then the MBMSE in Eq. 7 becomes [13] ∆2µ=σ 2 −σ 4I ˆϱ Q(µ)(D1) whereI ˆϱ Q(µ)is the QFI with respect toµof the displaced mixedstateˆϱ=e iµˆxˆ¯ρe−iµˆx. Thisshouldbedistinguished from the QFI ofˆρ(...

  6. [14]

    Quadrature measurement Suppose that we perform a quadrature measurement ˆponˆρ(µ)and obtain a measurement resultpwith a likelihood ofL(p|µ)∼ N(µ, 1√ 2 ). Then, the evidence is also Gaussianp(p)∼ N(0, ς)with zero mean and varianceς 2 = 1 2 +σ 2, and the posterior is Gaussian p(...

  7. [15]

    This means that the optimal estimator˜µn = 0 contains no additional information about the displace- mentµand thus the BMSE equals the prior variance σ2

    Number-resolving measurement In comparison, suppose that we perform a number- resolving measurementˆnand obtain a measurement re- sultnwith a Poissonian likelihood ofL(n|µ) =e − µ2 2 µ2n 2nn! such that the posterior distribution is p(µ|n) = µ2n σ2 + 1 n+ 1 2 2n+ 1 2 Γ n+ 1 2 σ...

  8. [16]

    Power and magnitude estimation To handle this issue with the sign of the displacement µ, suppose that we instead wish to estimate the power u=µ 2 or magnitudev=|µ|of the displacement. The Gaussian prior onµimplies that the power follows a scaled chi-squared distribution,π(u) =...

  9. [17]

    This limit represents the uninformative, uniform prior onR that is nevertheless normalised, i.e.∂ µπ(µ) = 0andR ∞ −∞dµ π(µ) = 1

    Wide-prior limit Returning to the case of estimatingµ, we now demon- strate that the wide-prior limit ofσ 2 ≫ 1 2 mentioned briefly above needs to be handled carefully. This limit represents the uninformative, uniform prior onR that is nevertheless normalised, i.e.∂ µπ(µ) = 0a...

  10. [18]

    We use the annihilation operator ˆbθ from Eq

    Extracting the temporal mode Let us start with the direct calculation without dis- cretisation. We use the annihilation operator ˆbθ from Eq. 46 to express the coherent state in Eq. 54 as follows |hθ⟩=e αθˆb† θ−α∗ θˆbθ |0⟩(E1) =e − 1 2 |αθ|2 eαθˆb† θ e−α∗ θˆbθ |0⟩ =e − 1 2 |αθ...

  11. [19]

    Let the anni- hilation operator of thejth time bin be ˆcj = 1√ δt Z (j+1)δt jδt dtˆa(t)(E5) where[ˆcj,ˆa†(t)] = 1√ δt χ(jδt,(j+1)δt) (t)and[ˆc j,ˆc† k] =δ jk forj, k∈Z

    Discretising the time domain Let us now instead discretise the bath of outgoing modes by coarse-graining them according to a short sam- pling timeδtsuch that the time domainRis broken into time bins(jδt,(j+t)δt)indexed byj∈Z. Let the anni- hilation operator of thejth time bin ...

  12. [20]

    Let us show this now for Gaussian and double-exponential window functions

    Alternate window functions We assumed a rectangular window function above but the results are similar with other window functions. Let us show this now for Gaussian and double-exponential window functions. Although we do not do this for the later amplitude and phase estimation...

  13. [21]

    Payne et al

    E. Payne et al. (in preparation), Prospects for post- merger remnant signal inference with photon-counting interferometry (2025)

    2025

  14. [22]

    P. D. Lasky, Gravitational wave astronomy, inMultimes- senger Astronomy in Practice, 2514-3433 (IOP Publish- ing, 2021) pp. 9–1 to 9–72

    2021

  15. [23]

    Sarin and P

    N. Sarin and P. D. Lasky, Multimessenger astronomy with a khz-band gravitational-wave observatory, Publ. Astron. Soc. Aust.39, e007 (2022)

    2022

  16. [24]

    J. E. Kim and G. Carosi, Axions and the strongCP problem, Rev. Mod. Phys.82, 557 (2010)

    2010

  17. [25]

    K. Choi, S. H. Im, and C. S. Shin, Recent progress in the physics of axions and axion-like particles, Annu. Rev. Nucl. Part. Sci.71, 225 (2021)

    2021

  18. [26]

    L. J. Rosenberg and K. A. Van Bibber, Searches for in- visible axions, Phys. Rep.325, 1 (2000)

    2000

  19. [27]

    P.W.Graham, I.G.Irastorza, S.K.Lamoreaux, A.Lind- ner, and K. A. van Bibber, Experimental searches for the axion and axion-like particles, Annu. Rev. Nucl. Part. Sci.65, 485 (2015)

    2015

  20. [28]

    R. D. Peccei, The strong cp problem and axions, inAx- ions: Theory, Cosmology, and Experimental Searches, edited by M. Kuster, G. Raffelt, and B. Beltrán (Springer Berlin Heidelberg, Berlin, Heidelberg, 2008) pp. 3–17

    2008

  21. [29]

    Battaglieri, A

    M. Battaglieri, A. Belloni, A. Chou, P. Cushman, B. Echenard, R. Essig, J. Estrada, J. L. Feng, B. Flaugher, P. J. Fox, et al., Us cosmic visions: new ideas in dark matter 2017: community report, arXiv preprint arXiv:1707.04591 (2017)

    2017 arXiv

  22. [30]

    Kubo, The fluctuation-dissipation theorem, Rep

    R. Kubo, The fluctuation-dissipation theorem, Rep. Prog. Phys.29, 255 (1966)

    1966

  23. [31]

    Buonanno and Y

    A. Buonanno and Y. Chen, Signal recycled laser- interferometer gravitational-wave detectors as optical springs, Phys. Rev. D65, 042001 (2002)

    2002

  24. [32]

    C. Bond, D. Brown, A. Freise, and K. A. Strain, Interfer- ometer techniques for gravitational-wave detection, Liv- ing Rev. Relativ.19, 1 (2016)

    2016

  25. [33]

    Macieszczak, M

    K. Macieszczak, M. Fraas, and R. Demkowicz- Dobrzański, Bayesian quantum frequency estimation in presence of collective dephasing, New Journal of Physics 16, 113002 (2014)

    2014

  26. [34]

    Jarzyna and R

    M. Jarzyna and R. Demkowicz-Dobrzański, True preci- sionlimitsinquantummetrology,NewJournalofPhysics 17, 013010 (2015)

    2015

  27. [35]

    Rubio and J

    J. Rubio and J. Dunningham, Quantum metrology in the presence of limited data, New Journal of Physics21, 043037 (2019)

    2019

  28. [36]

    Kaubruegger, D

    R. Kaubruegger, D. V. Vasilyev, M. Schulte, K. Ham- merer, and P. Zoller, Quantum variational optimization of ramsey interferometry and atomic clocks, Physical re- view X11, 041045 (2021)

    2021

  29. [37]

    C. D. Marciniak, T. Feldker, I. Pogorelov, R. Kaubrueg- ger, D. V. Vasilyev, R. van Bijnen, P. Schindler, P. Zoller, R. Blatt, and T. Monz, Optimal metrology with pro- grammable quantum sensors, Nature603, 604 (2022)

    2022

  30. [38]

    Direkci, R

    S. Direkci, R. Finkelstein, M. Endres, and T. Gefen, Heisenberg-limited bayesian phase estimation with low-depth digital quantum circuits, arXiv preprint arXiv:2407.06006 (2024)

    2024

  31. [39]

    Tsang, H

    M. Tsang, H. M. Wiseman, and C. M. Caves, Fundamen- tal Quantum Limit to Waveform Estimation, Phys. Rev. Lett.106, 090401 (2011)

    2011

  32. [40]

    H. Miao, R. X. Adhikari, Y. Ma, B. Pang, and Y. Chen, Towards the fundamental quantum limit of linear mea- surements of classical signals, Phys. Rev. Lett.119, 050801 (2017)

    2017

  33. [41]

    J. W. Gardner, T. Gefen, S. A. Haine, J. J. Hope, and Y. Chen, Achieving the fundamental quantum limit of linear waveform estimation,Phys. Rev.Lett.132, 130801 (2024)

    2024

  34. [42]

    A. S. Holevo,Probabilistic and Statistical Aspects of Quantum Theory(Springer Science & Business Media, 2011)

    2011

  35. [43]

    D. V. Vasilyev, A. Shankar, R. Kaubruegger, and P. Zoller, Optimal multiparameter metrology: The quan- tum compass solution, arXiv preprint arXiv:2404.14194 (2024)

    2024 arXiv

  36. [44]

    Monras, Phase space formalism for quantum estima- tion of Gaussian states (2013), arxiv:1303.3682 [quant- ph]

    A. Monras, Phase space formalism for quantum estima- tion of Gaussian states (2013), arxiv:1303.3682 [quant- ph]

    2013 arXiv

  37. [45]

    H. L. Van Trees,Detection, Estimation, and Modulation Theory: Part IV(John Wiley & Sons, Incorporated, 2002)

    2002

  38. [46]

    R. D. Gill and B. Y. Levit, Applications of the van Trees inequality: a Bayesian Cramér-Rao bound, Bernoulli , 59 (1995)

    1995

  39. [47]

    M. G. Genoni, M. G. Paris, G. Adesso, H. Nha, P. L. Knight, and M. Kim, Optimal estimation of joint pa- rameters in phase space, Physical Review A—Atomic, Molecular, and Optical Physics87, 012107 (2013)

    2013

  40. [48]

    J. W. Gardner, T. Gefen, S. A. Haine, J. J. Hope, J. Preskill, Y. Chen, and L. McCuller, Stochastic wave- form estimation at the fundamental quantum limit, PRX Quantum6, 030311 (2025). 29

    2025

  41. [49]

    J. W. Gardner, S. A. Haine, J. J. Hope, Y. Chen, and T. Gefen, Lindblad estimation with fast and pre- cise quantum control, arXiv preprint arXiv:2501.03364 (2025)

    2025

  42. [50]

    Rife and R

    D. Rife and R. Boorstyn, Single tone parameter estima- tion from discrete-time observations, IEEE Trans. Inf. Theory20, 591 (1974)

    1974

  43. [51]

    Steinhardt and C

    A. Steinhardt and C. Bretherton, Thresholds in fre- quency estimation, inICASSP 85 IEEE Int. Conf. Acoust. Speech Signal Process., Vol. 10 (1985) pp. 1273– 1276

    1985

  44. [52]

    James, B

    B. James, B. D. Anderson, and R. C. Williamson, Char- acterization of threshold for single tone maximum likeli- hood frequency estimation, IEEE transactions on signal processing43, 817 (2002)

    2002

  45. [53]

    Knockaert, The Barankin bound and threshold behav- ior in frequency estimation, IEEE Trans

    L. Knockaert, The Barankin bound and threshold behav- ior in frequency estimation, IEEE Trans. Signal Process. 45, 2398 (1997)

    1997

  46. [54]

    Schmitt, T

    S. Schmitt, T. Gefen, F. M. Stürner, T. Unden, G. Wolff, C. Müller, J. Scheuer, B. Naydenov, M. Markham, S. Pezzagna, et al., Submillihertz magnetic spectroscopy performed with a nanoscale quantum sensor, Science 356, 832 (2017)

    2017

  47. [55]

    Schmitt, T

    S. Schmitt, T. Gefen, D. Louzon, C. Osterkamp, N. Stau- denmaier, J. Lang, M. Markham, A. Retzker, L. P. McGuinness,andF.Jelezko,Optimalfrequencymeasure- ments with quantum probes, npj Quantum Information 7, 55 (2021)

    2021

  48. [56]

    Holevo, Estimation of shift parameters of a quantum state, Reports on Mathematical Physics13, 379 (1978)

    A. Holevo, Estimation of shift parameters of a quantum state, Reports on Mathematical Physics13, 379 (1978)

    1978

  49. [57]

    Chiribella, G

    G. Chiribella, G. D’ariano, and M. F. Sacchi, Opti- mal estimation of group transformations using entangle- ment, Physical Review A—Atomic, Molecular, and Op- tical Physics72, 042338 (2005)

    2005

  50. [58]

    H. M. Wiseman and G. J. Milburn,Quantum Measure- ment and Control(Cambridge University Press, 2009)

    2009

  51. [59]

    Direkci et al

    S. Direkci et al. (in preparation), Extending the dynamic range in atomic clocks with weak measurements (2025)

    2025

  52. [60]

    J. W. Gardner.boldBat. 2025.https://git.ligo.org/ jameswalter.gardner/boldbat

    2025

  53. [61]

    Demkowicz-Dobrzański, M

    R. Demkowicz-Dobrzański, M. Jarzyna, and J. Kołodyński, Quantum limits in optical interfer- ometry, Progress in Optics60, 345 (2015)

    2015

  54. [62]

    S. D. Bartlett, T. Rudolph, and R. W. Spekkens, Refer- ence frames, superselection rules, and quantum informa- tion, Reviews of Modern Physics79, 555 (2007)

    2007

  55. [63]

    Jeffreys, An invariant form for the prior probability in estimation problems, Proceedings of the Royal Society of London

    H. Jeffreys, An invariant form for the prior probability in estimation problems, Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences 186, 453 (1946)

    1946

  56. [64]

    Böttcher and S

    A. Böttcher and S. M. Grudsky,Toeplitz matrices, asymptotic linear algebra and functional analysis, Vol. 67 (Springer, 2000)

    2000

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.