{"id":"8cfe57d2-aca0-46e1-8192-90dc625b1061","arxiv_id":"2507.21736","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A coherently controlled superposition of a unitary and its inverse on a probe qubit gives optimal, axis-agnostic phase estimation with Fisher information 1.","lead":"A quantum sensor can estimate a rotation angle even when the rotation axis is completely unknown, using a helper qubit prepared in a coherent superposition to control the rotation and its inverse. The protocol achieves the optimal Fisher information of 1 without entanglement, relying only on coherence.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Agnostic claim depends on unproven access to U^{-1}; without a time-reversal oracle the controlled superposition in Eq. 5 cannot be built, so the agnostic advantage is conditional.","rationale":"I read the paper as claiming that coherence in the ancilla, via coherent control of U and U^{-1}, provides optimal axis-agnostic phase estimation without entanglement. The marginal ancilla calculation is straightforward and correct; the FI = 1 result is not in question. The load-bearing assumption is the physical availability of U^{-1}. The paper explicitly uses U^{-1} in Eq. 5 and references phase conjugation in footnote 25, but those references do not show that U^{-1} is obtainable under the stated 'no knowledge of the Hamiltonian' condition. In a black-box model, U^{-1} is additional resource. If the experimental platform cannot provide that resource without revealing n, the protocol is not agnostic in the same sense as [7], which uses only forward U and reveals H at the measurement stage. I also noticed a secondary issue: in Appendix B, Eq. (21) assigns cos²(τ/2)/2 to P_AP(+1, +1) and sin²(τ/2)(1+sin2θ sinφ)/2 to P_AP(+1, -1), but direct calculation gives both +-probe outcomes under ancilla '+' equal to cos²(τ/2)/2. This does not change the marginal probabilities in Eq. (9), so the main FI result stands, but the joint table should be corrected. The reader's CONDITIONAL verdict is appropriate; my concern is the same as the reader's weakest assumption.","tokens_in":13950,"tokens_out":23951,"duration_ms":289762,"concrete_test":"Test whether a fixed unitary V can satisfy V U V† = U^{-1} for all U ∈ SU(2). Check the two equations for U = e^{-iτ X/2} and U = e^{-iτ Y/2}; solving for V shows no solution, confirming that U^{-1} cannot be obtained from U by fixed gates alone. If no compilation exists, restate the central result as conditional on access to a time-reversal or phase-conjugation oracle, and adjust the comparison with [7] accordingly.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Equation (5) defines the protocol's central operation as a coherent controlled superposition of U and its inverse U^{-1}. Given access to L, the derivation of P± = cos²/sin²(τ/2) and FI = 1 is correct. However, the claim that this constitutes Hamiltonian-agnostic phase estimation rests on the assumption that U^{-1} can be implemented without knowing the rotation axis n. The paper's only justification (footnote [25]) invokes optical phase conjugation or reversing a control field. Those are physical resources requiring the ability to invert the dynamics, not just a black-box oracle for U; no fixed circuit of U and constant gates realizes U^{-1} for all U ∈ SU(2). If such a time-reversal oracle is unavailable, L cannot be realized and the protocol cannot start. This does not invalidate the algebra, but it narrows the claimed scope: the protocol needs an additional, unadvertised resource, and the comparison to protocols that use only forward U is not resource-fair.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a protocol for estimating the rotation angle τ of a single-qubit unitary U = e^{-iτ n·σ/2} without prior knowledge of the rotation axis n=(θ,φ). The protocol prepares an ancilla in the coherent state |+>, uses it to control whether the probe undergoes U or U^{-1} (the coherently controlled superposition L of Eq. (5)), and then measures the ancilla in the X basis. The outcome probabilities are P_± = cos²(τ/2) and sin²(τ/2), yielding Fisher information FI = 1 independent of (θ,φ). The paper computes the three-parameter Fisher information matrix for the joint Y⊗X measurement and shows it to be block-diagonal with F_{ττ}=1 and zero cross-terms with θ and φ, supporting the agnostic claim. Extensions to a maximally mixed probe and to tunable ancilla coherence are provided, together with a noise-robustness comparison against the entanglement-assisted protocol of Song et al. [7].","tokens_in":14141,"tokens_out":10554,"duration_ms":115273,"significance":"If the assumption of coherent access to U^{-1} is granted, the result is a clean and resource-light achievement: it shows that initial coherence, rather than initial entanglement or entangling measurements, suffices to saturate the quantum Fisher information bound for the rotation angle in an axis-agnostic way. The derivation is explicit, self-contained, and contains no fitted parameters; the predicted FI = 1 is falsifiable. The noise analysis in Appendix C gives a modest but concrete robustness advantage over the entangled protocol. The main limitation, discussed below, is the unadvertised resource of U^{-1}, which is load-bearing for the agnostic claim.","major_comments":[{"comment":"The protocol's central operation L requires access to both U and its inverse U^{-1}. The manuscript does not explain how to implement U^{-1} when the rotation axis n is unknown. The cited analogies—optical phase conjugation and reversing a control field—are specific physical mechanisms that are not guaranteed for an arbitrary unknown unitary and themselves presuppose partial knowledge of the dynamics. Without such a resource, the controlled superposition in Eq. (5) cannot be built, and the agnostic advantage disappears. This is a load-bearing issue because the paper's comparisons to protocols that use only forward U (standard phase estimation and the entanglement-assisted protocol of [7]) are not resource-fair. Please state explicitly that access to U^{-1} is an assumption, provide a constructive method able to produce U^{-1} from U with no axis knowledge, or reframe the claim as conditional on time-reversal capability.","section":"Coherently controlled superposition (CCS) enabled sensing, Eq. (5) and footnote [25]"},{"comment":"The Fisher information matrix in Eq. (22) is presented without a derivation. In particular, the off-diagonal block entries F_{θφ} and F_{φθ} are given in a compact form that is hard to verify. Since the block-diagonal structure (zero F_{τθ}, F_{τφ}) is the central quantitative claim, a derivation or a more transparent presentation showing how the (θ,φ) dependence cancels in the τ-row and τ-column would strengthen the paper. This is an accessibility issue rather than a correctness issue, but it makes the main result difficult for a reader to reproduce independently.","section":"Appendix B, Eq. (22)"}],"minor_comments":[{"comment":"The phrase \"More the optimality of the measurement, the better will be the improvement\" is awkward; consider rephrasing to \"A more optimal measurement yields a better estimate.\"","section":"Abstract"},{"comment":"The sentence \"This is a striking and distinctive result since it is the probe qubit which undergoes the phase encoding and the effect of it is witnessed on the ancilla–a clear manifestation of the coherent control mechanism –contrary to conventional phase estimation protocols\" is difficult to parse; please restructure and check the dash spacing.","section":"Introduction, near Eq. (9)"},{"comment":"The word \"irreverent\" appears where \"irrelevant\" is intended (e.g., \"irreverent to the knowledge\" and \"irreverent to the performance\"); please correct these typographical errors.","section":"Throughout"},{"comment":"The phrase \"F I= 1\" has spacing issues; it should be \"FI = 1\", and \"i.e,\" should be \"i.e.,\".","section":"Coherently controlled superposition section, after Eq. (9)"},{"comment":"The sentence \"the state of the ancilla acts like a fixed point in the interaction extracting all the information from a transformation (causal) it didn’t undergo\" is cryptic and would benefit from a concrete mathematical explanation of the fixed-point concept.","section":"Hindsight control section"},{"comment":"The caption describes the \"success probability of simulating the time-travel of the ancilla\"; since the time-travel narrative is interpretive, consider moving this phrase to the discussion and keeping the caption purely descriptive of the plotted quantity.","section":"Figure 4 caption"},{"comment":"The formula for the quantum Fisher information matrix for a pure state is correct, but the notation \"⟨∂iψ(λ)|ψ(λ)⟩\" is used without specifying that the derivatives are with respect to λ_i; a brief statement clarifying the partial-derivative convention would help.","section":"Appendix A, Eq. (12)"},{"comment":"Please add a sentence explaining how the Fisher information matrix in Eq. (22) is computed from the probabilities in Eq. (21), or provide a short derivation, to make the appendix self-contained.","section":"Appendix B, after Eq. (21)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript contains a correct and clean calculation, but the central claim of Hamiltonian-agnostic phase estimation hinges on the availability of U^{-1}, which is not justified. If the authors can provide a physical implementation of U^{-1} that works for arbitrary unknown n, the result would be much stronger; otherwise, the paper should be transparent about this resource assumption and soften the comparison to forward-U-only protocols. The time-travel narrative is decorative and could be moved to an outlook section without loss of content. The paper is within the scope of the journal and the result, with appropriate caveats, is publishable after revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nHere's the short version: the paper has one clean idea and does the algebra right, but the selling point—'agnostic' phase estimation—depends on unadvertised access to the inverse unitary. If you grant that oracle, the result is solid; if you don't, the protocol can't start.\n\nThe new thing is the controlled superposition L = |0><0|⊗U + |1><1|⊗U^{-1} used for single-qubit rotation angle estimation. The author shows that for ancilla prepared in |+>, measuring the ancilla in X gives P_± = cos²(τ/2), sin²(τ/2), so FI=1 independent of the rotation axis. That's correct, and it is genuinely simpler than the entanglement-based hindsight sensing of Song et al. The block-diagonal FI matrix in Appendix B supports the claim that τ decouples from θ, φ. The extension to a maximally mixed probe (FI=1 with no probe coherence) is a nice observation. The paper is also honest in comparing with [22] and [23]; the difference is clearly stated.\n\nThe soft spot is exactly where the stress-test note points. To build L, you need coherent control of U and U^{-1}. If U is a black box, no fixed circuit built from U and constant gates implements U^{-1} for all U∈SU(2). The footnote's appeal to optical phase conjugation or reversing a control field is a physical resource requirement: you need to invert the dynamics, which means you know something about the Hamiltonian (or have a time-reversal oracle). The paper does not flag this as an assumption. So the 'agnostic' claim should be narrowed to 'agnostic given access to U^{-1}'. That doesn't invalidate the algebra; it makes the comparison with forward-U-only protocols resource-unfair.\n\nA second, minor point: the noise robustness comparison in Appendix C compares depolarizing noise on the ancilla with a different noise model on the entangled state, so the 'slightly better' conclusion is not apples-to-apples. Not a big deal, but worth caveating.\n\nWho is this for? People working on quantum metrology, especially on channel-superposition and indefinite-causal-order approaches. They'll find it a clean, possibly useful protocol, and the time-travel narrative can be ignored. It deserves a serious referee: the idea is specific, the math is checkable, and the missing-resource issue is an editorial fix rather than a fatal flaw.\n\nMy recommendation: send it to review, but ask the author to state explicitly the assumption about U^{-1} availability at the outset and adjust the claims accordingly.","headline":"Neat controlled-superposition protocol for axis-agnostic phase estimation; the math is right, but the 'agnostic' claim rests on unstated access to U^{-1}.","tokens_in":14625,"tokens_out":2205,"would_cite":true,"duration_ms":24995,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that preparing an ancilla in $|+\\rangle$, using it to control whether a probe qubit undergoes an unknown rotation $U$ or its inverse $U^{-1}$, and measuring the ancilla in the $X$ basis, yields Fisher information…","keywords":["agnostic phase estimation","quantum metrology","Fisher information","quantum coherence","coherently controlled superposition","qubit rotation","unitary inversion","Hamiltonian-independent sensing"],"falsifier":"Prepare a qubit probe in $|0\\rangle$ and an ancilla in $|+\\rangle$, implement $L=|0\\rangle\\langle0|\\otimes U+|1\\rangle\\langle1|\\otimes U^{-1}$ for several axes with the same nominal $\\tau$, reveal the axes only after the run, and compare the ancilla counts to $P_+=\\cos^2(\\tau/2)$ and $P_-=\\sin^2(\\tau/2)$. If the outcome distribution or the estimated Fisher information depends on the axis, or if the Fisher information falls below 1 at $\\tau=\\pi/2$, the central claim is falsified.","tokens_in":13769,"feed_emoji":"🎯","tokens_out":9337,"duration_ms":103944,"temperature":0.7,"pith_summary":"This paper seeks to remove the main obstacle in qubit rotation-angle estimation: the need to know the rotation axis, equivalently the Hamiltonian, in advance to prepare an optimal probe and choose an optimal measurement. It claims that a single coherent ancilla qubit, controlling whether the probe undergoes an unknown rotation $U$ or its inverse $U^{-1}$, yields Fisher information $F=1$ for the rotation angle $\\tau$ for every axis, matching the known-H optimal bound without using entanglement. The ancilla outcome probabilities depend only on $\\tau$, so an experimenter can estimate the phase while remaining ignorant of the axis both before and after the interaction. A sympathetic reader would care because this is a resource-simple route to Hamiltonian-agnostic sensing in platforms where entanglement is costly, and it gives a concrete operational role to quantum coherence in saturating metrological precision.","feed_headline":"Coherent control reads a rotation angle without knowing its axis","feed_subtitle":"No axis calibration or entanglement is needed: one coherent ancilla delivers maximum Fisher information.","key_machinery":"The load-bearing object is the coherently controlled superposition (CCS) unitary $L=|0\\rangle\\langle0|_A\\otimes U+|1\\rangle\\langle1|_A\\otimes U^{-1}$, acting on a probe-ancilla product state with the ancilla in $|+\\rangle=(|0\\rangle+|1\\rangle)/\\sqrt{2}$ and the probe in $|0\\rangle$ or any fixed state. The ancilla coherence turns the controlled operation into an interference between forward evolution $U$ and backward evolution $U^{-1}$ on the probe; the axis dependence of the two paths cancels when the ancilla is measured in the $X$ basis, leaving outcome probabilities that depend only on $\\tau$. The $X$-basis measurement is called a coherence-preserving measurement, and it converts the ancilla's coherence into saturating Fisher information about the phase. The paper also tracks the resource quantitatively through the $l^1$-norm coherence $C(\\alpha)=\\sin 2\\alpha$ of the ancilla preparation.","core_discovery":"The paper's central claim is that a qubit rotation angle $\\tau$ can be estimated at the optimal Fisher-information level, $F(\\tau)=1$, without knowing the rotation axis $(\\theta,\\phi)$, by letting a coherent ancilla control whether the probe evolves under $U=e^{-i\\hat\\sigma\\cdot\\hat n\\tau/2}$ or under its inverse $U^{-1}$. With the ancilla prepared in $|+\\rangle$ and the probe in $|0\\rangle$, the controlled superposition $L=|0\\rangle\\langle0|_A\\otimes U+|1\\rangle\\langle1|_A\\otimes U^{-1}$ produces ancilla outcome probabilities $P_+=\\cos^2(\\tau/2)$ and $P_-=\\sin^2(\\tau/2)$, whose Fisher information with respect to $\\tau$ is exactly 1 for every axis. The three-parameter Fisher information matrix is block diagonal with $F_{\\tau\\tau}=1$ and zero correlations between $\\tau$ and the axis parameters, so the phase estimate is agnostic to axis knowledge. The result holds without entanglement in the input state or in the measurement, and the appendix shows that the same axis-agnostic $F_{\\tau\\tau}=1$ survives even when the probe starts maximally mixed, with the ancilla's coherence serving as the quantitative resource.","pith_inferences":["The paper leaves implicit that the practical meaning of 'agnostic' depends on the physical availability of $U^{-1}$; if a platform can implement phase conjugation, reversed control fields, or another unitary-inversion operation without knowing the axis, the scheme is immediately deployable, but the protocol itself does not explain how to realize $U^{-1}$ in a fully axis-agnostic setting.","Because the Fisher information matrix is singular in the axis block, the protocol is specifically tailored to estimating $\\tau$ alone; a natural extension is to ask whether higher-dimensional coherent control or multiple probes could extract axis information while retaining the same resource economy.","The appendix's formulas relating classical and quantum Fisher information to ancilla coherence suggest a direct experimental test: tune the ancilla superposition angle $\\alpha$, measure the ancilla $X$ statistics, and check that the Fisher information follows $C(\\alpha)^2\\sin^2\\tau/(1-C(\\alpha)^2\\cos^2\\tau)$, with full axis-agnostic saturation only at $\\alpha=\\pi/4$."],"forward_implications":["An experimenter who can implement $U$ and $U^{-1}$ under coherent control can estimate $\\tau$ at the standard quantum limit without calibrating the rotation axis, so the protocol works in fluctuating or uncharacterized fields.","Because the input is a product state and the final measurement is local on the ancilla, the scheme avoids both entangled state preparation and entangling measurements, making it cheaper than the hindsight and closed-timelike-curve based protocols it compares with.","The Fisher information about $\\tau$ is exactly 1 and decoupled from the axis parameters, so the Cramér-Rao bound $\\mathrm{Var}(\\hat\\tau)\\ge 1/N$ is saturated for any axis.","Even a maximally mixed probe suffices: the coherence lives only in the ancilla, and no dynamical entanglement is generated, so the advantage is attributable to coherent control rather than to entanglement.","Under depolarizing noise of strength $f$ on the ancilla, the Fisher information becomes $f^2\\sin^2\\tau/(1-f^2\\cos^2\\tau)$, which the paper reports as slightly better than the entanglement-based counterpart at the same noise levels."],"supporting_citations":[{"why":"Defines the entanglement-based agnostic phase estimation protocol that this paper's CCS protocol is compared against and claims to surpass without entanglement.","marker":"[7]"},{"why":"Supplies the standard quantum Fisher information bound $F(\\tau)\\le 1$ and the known-H optimality condition that the protocol removes.","marker":"[5]"},{"why":"Earlier protocol for parameter estimation through coherently superposed channels, from which this paper distinguishes its $U$ versus $U^{-1}$ superposition.","marker":"[22]"},{"why":"Coherence-based rotation-angle estimation via indefinite-time-direction dynamics, used as the comparison for the known-axis regime.","marker":"[23]"},{"why":"Defines the $l^1$-norm coherence measure used to quantify the ancilla resource and its relation to Fisher information.","marker":"[49]"},{"why":"States the relation between quantum coherence and quantum Fisher information used in the resource analysis.","marker":"[50]"}],"fun_headline_variants":["Coherent ancilla measures rotation angle with no axis knowledge","Axis-blind rotation sensing via coherent control","Optimal Fisher info for unknown-axis rotation","Coherence enables axis-agnostic rotation estimation","Coherently controlled ancilla gives optimal rotation sensing"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The protocol assumes that the inverse of the unknown rotation, $U^{-1}$, can be applied under coherent control without first knowing the rotation axis; if $U^{-1}$ cannot be physically realized without axis knowledge, the claimed agnostic advantage does not survive.","fun_headline_variants_meta":{"raw":{"variants":["Coherent ancilla measures rotation angle with no axis knowledge","Axis-blind rotation sensing via coherent control","Optimal Fisher info for unknown-axis rotation","Coherence enables axis-agnostic rotation estimation","Coherently controlled ancilla gives optimal rotation sensing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001101,"raw_usage":{"total_tokens":4654,"prompt_tokens":1065,"completion_tokens":3589,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":681,"completion_tokens_details":{"reasoning_tokens":3518}},"tokens_in":681,"tokens_out":3589,"duration_ms":31514,"temperature":1.0,"reasoning_tokens":3518,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T12:25:59.469334+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Prepare a qubit probe in $|0\\rangle$ and an ancilla in $|+\\rangle$, implement $L=|0\\rangle\\langle0|\\otimes U+|1\\rangle\\langle1|\\otimes U^{-1}$ for several axes with the same nominal $\\tau$, reveal the axes only after the run, and compare the ancilla counts to $P_+=\\cos^2(\\tau/2)$ and $P_-=\\sin^2(\\tau/2)$. If the outcome distribution or the estimated Fisher information depends on the axis, or if the Fisher information falls below 1 at $\\tau=\\pi/2$, the central claim is falsified.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the entanglement-based agnostic phase estimation protocol that this paper's CCS protocol is compared against and claims to surpass without entanglement."},{"cited_title":"Tóth and I","cited_arxiv_id":null,"evidence_quote":"Supplies the standard quantum Fisher information bound $F(\\tau)\\le 1$ and the known-H optimality condition that the protocol removes."},{"cited_title":"Altorio, M","cited_arxiv_id":null,"evidence_quote":"Earlier protocol for parameter estimation through coherently superposed channels, from which this paper distinguishes its $U$ versus $U^{-1}$ superposition."},{"cited_title":"Chapeau-Blondeau, Quantum parameter estimation on co- herently superposed noisy channels, Physical Review A 104, 032214 (2021)","cited_arxiv_id":null,"evidence_quote":"Coherence-based rotation-angle estimation via indefinite-time-direction dynamics, used as the comparison for the known-axis regime."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the $l^1$-norm coherence measure used to quantify the ancilla resource and its relation to Fisher information."},{"cited_title":"Baumgratz, M","cited_arxiv_id":null,"evidence_quote":"States the relation between quantum coherence and quantum Fisher information used in the resource analysis."}],"review_version":1}