{"id":"d2ce9635-22dd-44f6-8067-fd231e2e84f9","arxiv_id":"2501.00189","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Quadratic quantum metrology under collective dephasing reaches N^(-5/4) precision with product states and N^(-3/2) with spin-squeezed states plus nonlinear readout in the Zeno regime, claimed optimal.","lead":"This paper derives the best possible precision, as a function of sensor number N, for quadratic (nonlinear) frequency estimation when all sensors are affected by the same collective dephasing noise. It shows that temporally correlated noise lets even classical product states beat the standard linear quantum limit, and a practical spin-squeezed state with nonlinear readout reaches N^(-3/2) precision, matching an idealized maximally entangled state.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The N^-3/2 optimality claim rests on an HP approximation at <a†a> ~ J/2 plus a deferred no-go proof; both weaken the central claim.","rationale":"The reader's weakest-assumption analysis correctly identifies the quantitative faithfulness of the Holstein-Primakoff mapping for the perfect-echo OATS parameters as the most load-bearing assumption. The paper's own Eq. (D15) and Sec. IIIB acknowledge that the mean excitation number is of order J for these parameters, placing the state at the fringe of the HP regime. Since the headline N^{-3/2} achievability is derived entirely within this approximation, any uncontrolled correction could change the exponent and invalidate the primary result. I also flag the explicit gap between the abstract's claim of a proof of asymptotic optimality and Sec. V's statement that no formal proof is pursued; this is an internal inconsistency about the central claim, but it is secondary to the HP faithfulness issue because a proof of the no-go would be needed even if the achievability scaling survives. Independent support for other parts of the paper is real: the Phi-state results are exact, the CSS results are based on exact moment formulas, and the ratio-estimator construction is explicit and numerically benchmarked. These strengths do not, however, cover the OATS N^{-3/2} claim, which is the novel and advertised result. The proposed concrete test, exact finite-N simulation for increasing N, would settle whether the HP-derived exponent is a real feature of the spin dynamics or an artifact of the approximation; if it survives, the remaining issue is the missing optimality proof, which the authors should either provide or remove from the abstract.","tokens_in":33653,"tokens_out":6346,"duration_ms":61914,"concrete_test":"Exact diagonalization of the spin model in Eq. (1) with k=2 for N=16,32,64,128: prepare the PE OATS with mu=(2J)^-1/2, beta=-pi/2, evolve under collective dephasing with kappa(t)=kappa0^2 (omega_c t)^2, compute the time-optimized QFI (or the covariance of the echo readout), and fit the precision exponent. If the fitted exponent deviates from -3/2 outside finite-N extrapolation error, the HP-based scaling is an artifact. As a secondary check, re-derive Eq. (20) keeping the leading sqrt(1-a-dagger a/(2J)) correction; if the QFI acquires additional J-scaling terms, the PE result is not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that a perfect-echo OATS (mu=(2J)^-1/2, beta=-pi/2) achieves the asymptotically optimal N^{-3/2} scaling under collective dephasing in the Zeno regime. This claim has two load-bearing supports, and both are soft. First, the achievability is derived in the Holstein-Primakoff (HP) bosonic approximation, but Eq. (D15) gives <a-dagger a> = (1/4)[delta^{-1}+delta(1+4J^2 eta^2)+4J kappa(t)] ~ J/2 at t=0 for the PE parameters (delta=2J, eta=-(2J)^(-3/2)), i.e., an excitation fraction ~1/4 of the Bloch sphere radius. The paper itself states in Sec. IIIB that this state 'lies at the fringe of the HP regime of applicability.' The replacement sqrt(1-a-dagger a/(2J)) -> 1 used in the spin-boson mapping is then a ~13% correction at typical occupation, and the terms dropped from J_z = J - a-dagger a are of order J, so the effective quadratic boson Hamiltonian in Eq. (8) is not a controlled approximation. Second, the abstract asserts that the N^{-3/2} scaling 'we prove is asymptotically optimal,' but Sec. V explicitly says 'we do not pursue a formal proof here' and frames the matching no-go as a conjecture. Thus neither the achievability nor the optimality half of the headline result is fully established. If HP corrections alter the exponent, the main result fails; if the no-go cannot be proven, the word 'optimal' is unjustified.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes frequency estimation with quadratic signal encoding under classical collective dephasing with arbitrary temporal correlations, for N qubits described by H(t)=b J_z^2 + J_z ξ(t). It derives quantum Fisher information bounds for product coherent spin states and for a family of properly squeezed one-axis-twisted states (OATS), using a Holstein-Primakoff mapping and phase-space methods. The main scaling results are: Markovian collective dephasing limits all considered Gaussian states to N^{-1}; in the non-Markovian Zeno regime, CSS inputs reach N^{-5/4} and a perfect-echo OATS with an interaction-based readout reaches N^{-3/2}, the same scaling as a maximally entangled generalized GHZ (Phi) state. The paper also constructs a two-observable ratio estimator that is asymptotically unbiased under dephasing and claims that it reaches the same precision as the standard estimator. The abstract states that the N^{-3/2} scaling is 'proved asymptotically optimal.'","tokens_in":33961,"tokens_out":4936,"duration_ms":53243,"significance":"If the central claims are fully established, this is a valuable contribution to nonlinear quantum metrology under realistic dephasing. The analytic SLD and QFI derivation for Gaussian OATS in Sec. III B and Appendix D is detailed and self-contained, with the noiseless and exact Phi-state limits recovered consistently, and the numerical checks in Figs. 2 and 3 support the CSS scalings. The ratio estimator is a practically useful construction that addresses a real bias problem. However, the headline optimality result rests on two load-bearing points that are not yet established: the controlled validity of the Holstein-Primakoff approximation at the perfect-echo working point, and the matching upper bound that would justify the word 'optimal.'","major_comments":[{"comment":"The achievability of the N^{-3/2} scaling for the perfect-echo OATS relies on the Holstein-Primakoff Hamiltonian in Eq. (8), but the paper's own Eq. (D15) gives <a-dagger a> ≈ J/2 at t=0 for the PE parameters μ=(2J)^{-1/2}, β=-π/2, i.e. the excitation number is of order N rather than ≪J. The manuscript itself states in Sec. III B that this state 'lies at the fringe of the HP regime of applicability.' The replacement sqrt(1-a-dagger a/(2J))→1 drops corrections of order unity in J_z, so the effective bosonic Hamiltonian is not a controlled approximation at the exact working point used for the headline exponent. The illustrative phase-space plot in Fig. 1 uses N=10 and cannot validate the asymptotic scaling. I ask for a quantitative assessment, for example exact finite-J simulation of the spin dynamics for the PE protocol up to moderate N, or a rigorous bound on the error induced by the HP truncation in the QFI; without this, the N^{-3/2} achievability claim is not fully supported.","section":"Sec. III B 1 and Appendix D, Eq. (D15)"},{"comment":"The abstract claims that the N^{-3/2} scaling 'we prove is asymptotically optimal,' but Sec. V explicitly states 'we do not pursue a formal proof here' and frames the matching no-go as a conjecture based on ideas from Ref. [28], which is itself 'in preparation.' What the paper actually proves is that the PE OATS attains the same exponent as the Phi state. That is an equality with a known state, not an upper bound over all strategies. Please either supply the missing optimality proof or revise the abstract and Sec. V to state that N^{-3/2} is attained by the PE OATS and is conjectured, not proved, to be asymptotically optimal.","section":"Abstract and Sec. V"},{"comment":"The QFI optimization is performed using only the term F_Q^(A) in Eq. (20), and the text in Appendix D states that F_Q^(B) can be neglected for α<1/2, while for α=1/2 both terms scale equally. The perfect-echo OATS is precisely the boundary case α=1/2, so the claim that dropping F_Q^(B) 'does not otherwise change the main conclusions' needs an explicit demonstration at the PE parameters. This does not affect the claimed exponent, but it affects the constants in Table I and the statement that the readout in Eq. (22) saturates the QCRB.","section":"Sec. III B 2, Eq. (20)"}],"minor_comments":[{"comment":"The notation 'Ju,u ∈ {u,y,z}' contains a typo; it should read J_u with u ∈ {x,y,z}.","section":"Eq. (1)"},{"comment":"The sentence 'In the limit of temporally uncorrelated, Markovian noise is spectrum is flat' should read 'the noise spectrum is flat.'","section":"Sec. II A"},{"comment":"'inducediffusion along the Jy-direction' is missing a space and should read 'induces diffusion.'","section":"Fig. 1 caption"},{"comment":"The set K_hat{z} of properly squeezed states is used before being formally defined; please give an explicit definition in the main text rather than only in Appendix D.","section":"Sec. III B 1"},{"comment":"The extension of the ratio estimator to OATS states is described only as 'conceptually straightforward' and is not derived. Since the abstract and Sec. I state that the ratio estimator works 'without detriment to the achievable precision,' the scope of that claim should be clarified: a derivation or a caveat is needed for the entangled OATS case.","section":"Sec. IV B"},{"comment":"The optimality conjecture for the no-go bound relies on Ref. [28], which is listed as 'in preparation.' For a journal submission, the status of this reference should be made explicit, or the conjecture should be stated without depending on an unpublished work.","section":"Ref. [28]"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a substantial and mostly careful analytic derivation, and I think the CSS scaling results and the ratio estimator are solid enough for publication after revision. My main concern is that the two headline statements—'N^{-3/2} is achieved by the PE OATS' and 'N^{-3/2} is asymptotically optimal'—are both stronger than what is currently established. The achievability issue at <a-dagger a> ≈ J/2 is a genuine correctness risk that can be addressed by numerical checks, and the optimality issue is a matter of matching the claims to the proof provided. I would not recommend rejection, because the gaps are local and fixable within the scope of the manuscript."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Riberi and Viola have produced a genuinely useful piece of work, but the headline is stronger than what the paper actually establishes. The phase-space SLD machinery for Gaussian one-axis-twisted states under collective dephasing is the real contribution: it gives analytic QFI expressions, a clear readout prescription, and recovers the noiseless limits. The CSS Zeno scaling N^{-5/4} is new, well-derived, and backed by numerics. The ratio estimator for bias-free estimation is a practical add-on that appears to work. For those parts, the paper deserves serious attention.\n\nThe soft spots are the two load-bearing supports of the N^{-3/2} claim, and the authors essentially acknowledge both. First, the abstract says the scaling \"we prove is asymptotically optimal,\" but Sec. V defers the matching no-go, calling it a conjecture and offering no proof. That is a real mismatch between abstract and body. Second, the perfect-echo OATS parameters that give N^{-3/2} put the mean excitation number at about J/2, i.e., <a†a> ~ N/4. The HP mapping requires <a†a> ≪ J. The paper says the state \"lies at the fringe of the HP regime,\" which is an honest caveat, but it means the N^{-3/2} exponent is not established beyond the approximate bosonized dynamics. The exact Phi-state benchmark shows that N^{-3/2} is attainable in principle; the question is whether the OATS protocol actually attains it in the full spin model. That question is left open.\n\nNeither of these problems invalidates the CSS results or the QFI method. But they do mean the central \"optimal\" claim is currently a conjecture plus an approximation. A referee should ask for a clearer statement of what is proven versus conjectured, and for either a better justification of the HP regime or a numerical check against the spin model for moderate N.\n\nThis paper is worth engaging with. Send it to review. It needs revision, not rejection. The authors are technically skilled and the derivations are transparent enough that a good referee can make the paper much more honest.","headline":"The CSS scaling and the QFI machinery are solid, but the N^{-3/2} optimality claim is an acknowledged conjecture sitting at the edge of the HP approximation, so the abstract oversells it.","tokens_in":34543,"tokens_out":3363,"would_cite":true,"duration_ms":32001,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Under collective dephasing, a squeezed one-axis-twisted state with an echo readout reaches the same N^-3/2 quadratic frequency-estimation precision as a maximally entangled GHZ state.","keywords":["quantum metrology","nonlinear metrology","collective dephasing","one-axis-twisted states","quantum Fisher information","Holstein-Primakoff mapping","frequency estimation","Zeno regime"],"falsifier":"Simulate the full N-qubit spin dynamics without the Holstein-Primakoff approximation for the perfect-echo state $\\mu=(2J)^{-1/2}$, $\\beta=-\\pi/2$ under collective dephasing in the Zeno regime, optimizing over the encoding time near $\\tau_{\\rm opt}=1/(2\\kappa_0\\omega_c)|\\csc\\theta_{\\rm opt}|(N\\delta/2)^{-1/2}$ with $\\theta_{\\rm opt}=\\arccos\\sqrt{2/3}$; if the time-optimized variance scales worse than $N^{-3/2}$ (for instance $N^{-5/4}$) as $N$ grows, the central claim collapses. A cold-atom implementation measuring $J_y$ after the anti-squeezing echo would provide the same check experimentally.","tokens_in":33415,"feed_emoji":"⚛️","tokens_out":11970,"duration_ms":127281,"temperature":0.7,"pith_summary":"This paper asks how much precision a quadratic (nonlinear) frequency sensor can retain when all N qubit probes suffer identical classical dephasing, with arbitrary temporal correlations in the noise. It establishes that under Markovian dephasing every state considered — product, squeezed, or maximally entangled — is capped at the linear Heisenberg scaling $N^{-1}$, so entanglement provides no advantage. With temporally correlated noise in the short-time Zeno regime, a product coherent spin state already reaches $N^{-5}$/4, while a properly squeezed one-axis-twisted state followed by an anti-squeezing echo readout reaches $N^{-3}$/2, matching a generalized GHZ state. The paper proves this last bound is achievable and argues, through a conjecture deferred to the conclusion, that $N^{-3}$/2 is asymptotically optimal. It also constructs a two-observable ratio estimator that removes noise-induced bias at twice the measurement cost without sacrificing the quantum Cramér-Rao precision.","feed_headline":"Squeezed states match GHZ precision scaling under dephasing","feed_subtitle":"It matches the GHZ-state N^-3/2 scaling using a squeezed state despite collective dephasing.","key_machinery":"The load-bearing object is the family $\\mathcal K_{\\hat z}$ of Gaussian (properly squeezed) one-axis-twisted states $|{\\rm OATS}\\rangle=e^{-i\\beta J_z}e^{-i\\mu J_x^2}|{\\rm CSS}\\rangle$ in the Holstein-Primakoff low-excitation regime, where $J_+\\simeq\\sqrt{2J}\\,a$, $J_z=J-a^\\dagger a$, and the noisy Hamiltonian becomes a driven oscillator with quadratic position coupling $H\\simeq b(\\sin\\theta\\,\\hat x+J\\cos\\theta)^2+(\\sin\\theta\\,\\hat x+J\\cos\\theta)\\xi(t)$. In this representation the Wigner function remains Gaussian, the symmetric logarithmic derivative is exactly solvable as a polynomial at most quadratic in $\\hat x,\\hat p$, and the quantum Fisher information splits as $F_Q=F_Q^{(A)}+F_Q^{(B)}$ with a vanishing cross term. The leading contribution is controlled by the effective squeezing parameter $\\delta=\\cos^2\\beta+(1+4J^2\\mu^2)\\sin^2\\beta+2J\\mu\\sin(2\\beta)$, and the associated optimal POVM is the anti-squeezed (echo) observable $e^{i\\eta J_x^2/2}J_y e^{-i\\eta J_x^2/2}$, which avoids single-particle resolution.","core_discovery":"By mapping the spin dynamics to a bosonic mode through the Holstein-Primakoff transformation and solving the symmetric-logarithmic-derivative equation in phase space, the paper obtains an exact asymptotic quantum Fisher information for properly squeezed one-axis-twisted states under collective dephasing. For the optimal signal direction $\\theta_{\\rm opt}=\\arccos\\sqrt{2/3}$ and Zeno-regime noise $\\kappa(t)\\simeq\\kappa_0^2(\\omega_c t)^2$, the time-optimized precision is $\\Delta\\hat b_{\\rm opt}^{\\rm OATS}\\propto\\delta^{-1/4}N^{-5/4}$, where $\\delta$ is the effective quadrature-squeezing parameter. At the perfect-echo parameters $\\mu=(2J)^{-1/2}$, $\\beta=-\\pi/2$, one has $\\delta=2J$, which pushes the scaling to $\\Delta\\hat b_{\\rm opt}^{\\rm PE}\\propto N^{-3/2}$, the same exponent as the maximally entangled $|\\Phi\\rangle$ state. The saturating readout is the interaction-based echo observable $O=e^{i\\eta J_x^2/2}J_y e^{-i\\eta J_x^2/2}$, corresponding to the leading part of the SLD; the cross term with the quadratic part vanishes exactly for these states. The claimed asymptotic optimality of the $N^{-3/2}$ bound is stated in Sec. V as a conjecture, with the matching upper bound expected to follow from techniques used in the linear metrology setting.","pith_inferences":["If the deferred optimality proof is completed, N^-3/2 becomes a universal ceiling for quadratic encoding under one-body collective dephasing; a natural next test is whether two-axis countertwisting or higher-order squeezing can enter a different excitation regime and change the exponent.","Because the perfect-echo state sits at the fringe of the Holstein-Primakoff regime, exact finite-N spin simulations could determine whether the N^-3/2 exponent survives outside the Gaussian approximation or merely the prefactor changes.","The ratio-estimator construction should extend to entangled one-axis-twisted inputs through a second-order cumulant expansion, making bias-free estimation available for the same states that achieve the N^-3/2 scaling.","The results suggest a resource hierarchy for nonlinear metrology: temporal noise correlations are what unlocks sub-Heisenberg scaling, while squeezing substitutes for entanglement as the state resource."],"forward_implications":["If the N^-3/2 bound is asymptotically optimal, quadratic encoding under collective dephasing offers a super-Heisenberg advantage over the linear N^-1 limit that is accessible with experimentally available squeezed states rather than fragile GHZ states.","Under Markovian collective dephasing, none of the considered input states beats N^-1; any precision enhancement in that regime would have to come from reducing noise correlations, not from input entanglement or squeezing.","The state-independent optimal angle $\\theta_{\\rm opt}=\\arccos\\sqrt{2/3}$ gives a concrete tuning prescription for generalized Ramsey protocols with quadratic encoding.","The ratio estimator yields an asymptotically unbiased frequency estimate under dephasing at the cost of doubling the measurement resources, with no asymptotic loss in precision relative to the quantum Cramér-Rao bound.","Together the achievability result and the optimality conjecture imply the generalized no-go $N^{-(k-p/2)}$ for k-th order nonlinearities under p-body collective dephasing, so interactions can beat the linear Heisenberg limit but cannot beat the generalized classical bound."],"supporting_citations":[{"why":"Supplies the Holstein-Primakoff phase-space machinery, including the Mehler-kernel representation of squeezed states, that the paper adapts to non-perpendicular signal directions and noise.","marker":"[63]"},{"why":"Provides the perfect-echo OATS parameters $\\mu=(2J)^{-1/2}$, $\\beta=\\pi/2$ and the interaction-based readout that produce the $N^{-3/2}$ scaling.","marker":"[53]"},{"why":"Establishes the noiseless quadratic-metrology bounds (generalized SQL for product states and generalized HL for entangled states) that the noisy QFI must recover in the $\\kappa=0$ limit.","marker":"[37]"},{"why":"Defines one-axis-twisted spin-squeezed states and the minimal-dispersion parameters used for the KU OATS entry in Table I.","marker":"[65]"},{"why":"Supplies the spectral-overlap expression for the dephasing coefficient and the two-observable ratio-estimator strategy that the paper generalizes to quadratic encoding.","marker":"[33]"},{"why":"Provides the cumulant-expansion and noise-bias framework used for the CSS ratio estimator and the operating-point analysis.","marker":"[34]"},{"why":"Establishes the non-Markovian/Zeno regime as the setting in which scaling advantages over Markovian noise can appear.","marker":"[27]"},{"why":"Supplies the generalized no-go notation $N^{-(k-p/2)}$ and the result that uncorrelated dephasing preserves nonlinear scaling, which the paper's conjecture extends to collective one-body dephasing.","marker":"[49]"},{"why":"Identifies the Zeno limit in frequency estimation with non-Markovian environments that defines the short-time regime used throughout the paper.","marker":"[52]"}],"fun_headline_variants":["Squeezed states hit GHZ-level precision under dephasing","Nonlinear metrology: squeezed states achieve N^-3/2 under dephasing","Squeezed states reach optimal quantum scaling despite dephasing","Squeezed sensors match entanglement-limited precision bounds","Squeezed states beat classical scaling in dephased metrology"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the Holstein-Primakoff low-excitation mapping stays quantitatively faithful for the perfect-echo one-axis-twisted parameters, even though Eq. (D15) shows the mean excitation number is already of order N and the paper itself places that state at the fringe of the mapping's validity.","fun_headline_variants_meta":{"raw":{"variants":["Squeezed states hit GHZ-level precision under dephasing","Nonlinear metrology: squeezed states achieve N^-3/2 under dephasing","Squeezed states reach optimal quantum scaling despite dephasing","Squeezed sensors match entanglement-limited precision bounds","Squeezed states beat classical scaling in dephased metrology"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000396,"raw_usage":{"total_tokens":2157,"prompt_tokens":1111,"completion_tokens":1046,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":727,"completion_tokens_details":{"reasoning_tokens":956}},"tokens_in":727,"tokens_out":1046,"duration_ms":10405,"temperature":1.0,"reasoning_tokens":956,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:57:06.410284+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the full N-qubit spin dynamics without the Holstein-Primakoff approximation for the perfect-echo state $\\mu=(2J)^{-1/2}$, $\\beta=-\\pi/2$ under collective dephasing in the Zeno regime, optimizing over the encoding time near $\\tau_{\\rm opt}=1/(2\\kappa_0\\omega_c)|\\csc\\theta_{\\rm opt}|(N\\delta/2)^{-1/2}$ with $\\theta_{\\rm opt}=\\arccos\\sqrt{2/3}$; if the time-optimized variance scales worse than $N^{-3/2}$ (for instance $N^{-5/4}$) as $N$ grows, the central claim collapses. A cold-atom implementation measuring $J_y$ after the anti-squeezing echo would provide the same check experimentally.","supporting_citations":[],"review_version":1}