{"id":"33dbd1ce-5fb7-42ed-9e75-21146caf48e2","arxiv_id":"2510.09216","paper_version":2,"verdict":"REJECT","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":2,"one_line_summary":"An indefinite-time-direction encoding creates a photon OAM superposition whose generator variance grows as T², yielding a 1/T² metrological scaling that is presented as resource-free.","lead":"A quantum-switch protocol is claimed to beat the standard Heisenberg precision limit, reaching a 1/T² scaling for rotation sensing using photon orbital angular momentum. The paper's own equations show the generator uncertainty grows as T², so the advertised 'no extra probe resources' claim does not hold under standard resource accounting.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Resource-free claim depends on a nonstandard decomposition-averaged variance; Eq. (4) shows the canonical-momentum variance grows as T^2, making the scheme equivalent to an OAM-superposition probe.","rationale":"The paper's algebraic derivations appear internally consistent: Eq. (4), Eq. (6), and the experimental probabilities Eq. (10)/Eq. (13) correctly follow from the stated unitaries. The experiment also demonstrates the predicted interference fringes. The load-bearing problem is interpretive: the abstract claims a nonlinear precision enhancement 'without using probe-side information resources' and 'without relying on increasingly informative probe states.' That claim is supported only by measuring resources as the decomposition-averaged uncertainty of V_S on the reduced probe state, a quantity that assigns zero variance to each OAM branch and therefore hides the coherence between branches. The standard resource in quantum metrology — and the quantity appearing in the paper's own Eq. (1) — is the variance of the parameter generator on the actual sensing state. Eq. (4) shows this variance contains T_C^2 T_S^2, so the scheme does use a generator variance growing as T^2. The concrete test settles the issue: if the ITD protocol's statistics are exactly reproduced by a standard rotated OAM superposition, then the nonlinear scaling is not a new resource-free effect but a known Heisenberg-limited strategy with a state-preparation gadget. The reader's verdict of rejection is therefore appropriate, though the mathematical core and experiment retain value as a control-enhanced metrology demonstration.","tokens_in":28926,"tokens_out":9875,"duration_ms":393913,"concrete_test":"Re-express the protocol as a standard metrology circuit: take initial probe state |ψ_T⟩ = (|+T⟩+|−T⟩)/√2, unitary e^{-igT L_z}, and measure in the basis (|+T⟩±|−T⟩)/√2. Compute the probabilities and Fisher information. If they reproduce Eq. (10) and F(g)=4T^4, the ITD generating process is equivalent to state preparation with large generator variance, and the resource-free enhancement claim fails. Also compute Var(K̅_I) on the original joint input using the standard variance (not decomposition-averaged); if it equals T^2, the paper's own Eq. (1) already assigns the scheme a growing generator variance.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim — that nonlinear precision scaling is achieved without probe-side information resources — rests entirely on quantifying resources by ΔV̄_S, the decomposition-averaged uncertainty of V_S on the reduced probe state (Supplemental Note 2). Under this measure, each OAM branch |±m⟩ has zero variance of L_z, so the average is zero. But the paper's own precision bound, Eq. (1), is set by ΔK̄, the variance of the canonical momentum on the initial sensing state. Eq. (4) gives K̅_I = (K_S − T_C T_S)|0⟩⟨0| + (K_S + T_C T_S)|1⟩⟨1|, so for the ancilla in |+⟩ the variance of K̅_I is Var(K_S) + T_C^2 T_S^2. With T_C = T_S = T this grows as T^2, exactly the 'increasingly large variance of the parameter generator' the abstract says is avoided. The decomposition-averaged measure discards the coherence between the two OAM branches that actually carries the phase information; the reduced probe alone has zero QFI, while the joint state has QFI 4T^4. Thus the claim 'without relying on increasingly informative probe states' is an artifact of the resource definition. Expressed in the standard variance measure, the protocol is equivalent to preparing the OAM superposition (|+T⟩+|−T⟩)/√2 before a rotation e^{-igT L_z}, which is a standard Heisenberg-limited probe with generator variance T^2.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a quantum metrology protocol in which an indefinite-time-direction (ITD) generating process, implemented by a two-level ancilla and a Hamiltonian conjugate to the parameterizing generator, is applied before the parameterizing unitary. The authors derive precision bounds δg ≥ 1/(2√ν√(ΔV_S²T²+T⁴)) for a single pass and δg ≥ 1/(√ν(N²+N)) for N sequential passes, and report an optical experiment using OAM and Q-plates that confirms the predicted probabilities and RMSE scalings. The central claim is that this nonlinear scaling is achieved 'without using probe-side information resources,' i.e., without increasingly informative probe states.","tokens_in":29263,"tokens_out":14343,"duration_ms":125612,"significance":"If the resource-free claim were correct, the result would overturn the standard resource accounting for Heisenberg-limited metrology: a fixed two-level ancilla and noncommuting unitaries would beat the usual 1/T and 1/N limits without paying the conventional probe-side costs. The mathematical derivation of the scaling for the specific protocol appears internally consistent, and the experimental data match the theoretical curves. The paper also offers a unified uncertainty-principle reformulation of metrological bounds, which is a useful perspective. However, the advertised resource-free character of the enhancement does not survive standard resource accounting, and the experimental comparison is not made against a resource-equivalent standard scheme. The significance of the result therefore depends on accepting a nonstandard and, in my view, unjustified resource measure.","major_comments":[{"comment":"The 'no probe-side information resources' claim rests on the decomposition-averaged uncertainty ΔV̄_S of V_S on the reduced probe state. This measure discards the variance between the two OAM branches: after the Q-plate, the reduced probe is approximately (1/2)(|ψ(+m)⟩⟨ψ(+m)| + |ψ(-m)⟩⟨ψ(-m)|), whose standard variance of L_z is m² for initial OAM 0, while the branch-averaged variance computed in Supp. Note 2 remains the initial small value. Equation (S7) shows that the joint-state variance of the canonical momentum grows as ΔV_S²T² + T⁴. Thus, under the standard variance measure used in the paper's own Eq. (1), the protocol is equivalent to preparing an OAM superposition with generator variance T² before a rotation e^{-igT L_z}; the QFI is 4T⁴ and δ ∝ 1/T² is the standard Heisenberg limit for that probe. The resource-free claim is therefore an artifact of the resource definition.","section":"Supplemental Note 2; Eqs. (4) and (S7)"},{"comment":"The experimental comparison is not resource-equivalent. The dashed 'linear Heisenberg limit' is drawn for ΔV_S=1, i.e., a probe with small L_z variance. A standard scheme using a fixed OAM superposition (|+T⟩+|-T⟩)/√2 and the same total evolution time T has ΔL_z=T and precision δ=1/(2√ν T²), exactly the solid line. The observed scaling therefore does not demonstrate an advantage over a standard probe with comparable resources. Moreover, for the N-fold protocol the readout requires an N-order Q-plate (Fig. 3, 'N×m order'); this is an additional resource that scales with N and is not accounted for in the resource count.","section":"Eqs. (10)-(11), Figs. 4-5"},{"comment":"The Discussion states that the enhancement 'arises from the nonlinearly increased phase uncertainty on the ancilla.' This is an explicit acknowledgment that the ancilla—part of the sensing system—acquires a phase uncertainty growing as T², i.e., an increasingly informative probe-side state. This directly contradicts the abstract's claim that the protocol works 'without relying on increasingly informative probe states.' The resource argument is therefore internally inconsistent.","section":"Discussion, third paragraph of Sec. III"}],"minor_comments":[{"comment":"The average uncertainty ΔK̄ is defined via an arbitrary decomposition ρ=Σp_i|ψ_i⟩⟨ψ_i|. Please specify whether this is the spectral decomposition; otherwise the bound is not basis-independent and the resource measure is ill-defined.","section":"Eq. (1)"},{"comment":"There is a typo in the title as it appears on the arXiv: 'Metr ology' should be 'Metrology'.","section":"Title"},{"comment":"The error bars are not defined in the captions. Please state how the RMSE error bars are computed from the 20 groups of 30 estimates.","section":"Figs. 4 and 5"},{"comment":"The derivation uses the formal commutator [H_C,V_S]=i, while the experiment implements the unitary e^{±imφ}. The relationship between the formal calculation and the unitary implementation should be clarified, including the domain caveat of the angle operator, beyond the brief citation of Ref. [51].","section":"Sec. II B, after Eq. (4)"}],"recommendation":"reject","confidential_remarks":"The core problem is the resource accounting. The authors could potentially reframe the work as a demonstration of 1/T² scaling using OAM superpositions and a quantum switch, but the title and abstract claim of 'without probe-side information resources' is not defensible under standard metrological resource measures. The mathematical derivation and the experiment are sound in themselves, but the advertised conclusion is not supported. If the journal values the experimental technique, a major revision with a more honest resource comparison might be considered; as it stands, the overreach is load-bearing."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the quadratic scaling in T and N is genuinely in the paper—the derivations check out and the experiment matches the predicted probabilities. The problem is the headline claim that this comes 'without using probe-side information resources.' That claim doesn't survive a careful look at their own definitions. Eq. (4) shows the canonical momentum uncertainty grows as T² (variance T⁴), and the QFI of the joint state before parameterization is 4T⁴. The probe that actually enters the parameterization is the OAM superposition (|+T>+|−T>)/√2 entangled with polarization, which has generator variance T². The 'bounded ΔV̄_S' constraint they impose on the reduced probe state discards the coherence between the two OAM branches—the very coherence that carries the phase. A standard scheme using that same superposition and the same total time reaches the same precision. So the enhancement is not resource-free; the resource is a large generator variance, prepared by the Q-plate rather than put into the initial state. The paper's own Discussion admits this when it says the gain comes from 'nonlinearly increased phase uncertainty on the ancilla'—which contradicts the abstract.\n\nWhat is genuinely new: the sequential-N extension (δ ∝ 1/N²) and the specific optical implementation with Q-plates and Dove prisms. The experimental work is solid: probabilities, RMSE, and error bars align with theory, and the coincidence-counting procedure is described carefully. The theoretical framework relating precision to canonical momentum uncertainty is a clean reformulation of the QCRB, and the proof that ancilla-free schemes are limited to linear scaling under their own uncertainty constraint is fine. The single-shot version, however, was already published by the same group (ref 40), so the novelty rests mostly on the N-extension and the demonstration.\n\nSoft spots: the resource claim is load-bearing and not justified; the discussion of prior super-Heisenberg work is a bit hand-wavy (e.g., the Berry-curvature argument for ICO is asserted, not derived); and the constraint on ΔV̄_S of the reduced probe is not the quantity that controls the precision. A fair comparison would give the standard scheme the same OAM superposition and total time, which yields the same precision—so the advertised 'enhancement' disappears.\n\nWho this is for: experimentalists in quantum optics and quantum-switch metrology will find the setup and data useful; it is also a good case study for a reading group on how resource definitions can obscure a scaling advantage. But as a claim about surpassing the Heisenberg limit without resources, it fails.\n\nRecommendation: send it to peer review rather than desk reject—the experiment and sequential protocol deserve scrutiny—but expect referees to demand substantial revision. If the 'resource-free' language is kept, I would not accept it.","headline":"The 1/T² and 1/N² scalings are real and the experiment is clean, but the 'no probe-side resources' claim is an accounting artifact: the generating process simply prepares a large-OAM superposition probe.","tokens_in":29796,"tokens_out":6243,"would_cite":true,"duration_ms":58713,"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":"A quantum switch that superposes forward and backward time evolution in the encoding step yields a precision limit that scales as the inverse square of time, not linearly.","keywords":["Quantum metrology","Heisenberg limit","Super-Heisenberg scaling","Indefinite time direction","Quantum switch","Canonical momentum uncertainty","Quantum Fisher information","Orbital angular momentum"],"falsifier":"For the OAM experiment, perform full tomography of the probe–ancilla state at T=1,2,3,4 and compute the quantum Fisher information F(g) of the parameter. The paper's own probabilities give F=4T⁴; if a measurement shows F growing as T⁴, the enhancement is carried by the generator variance on the sensing state, and the 'resource-free' status depends entirely on the paper's chosen measure on V_S.","tokens_in":28716,"feed_emoji":"🕰️","tokens_out":4849,"duration_ms":40637,"temperature":0.7,"pith_summary":"The paper argues that the usual Heisenberg limit—precision improving only linearly with interrogation time T or gate count N—is not a fundamental bound on quantum metrology once the metrological process itself can run in an indefinite time direction. It reformulates the precision limit as 1/(2√ν ΔK̄), where ΔK̄ is the average uncertainty of the canonical momentum conjugate to the unknown parameter, and shows that a generating process implemented with a quantum switch (a coherent superposition of forward and backward time evolution) converts noncommutativity between the generating and parameterizing operations into a quadratic growth of ΔK̄. The result is a nonlinear scaling δg ∝ 1/T² for a single pass and δg ∝ 1/N² for N sequential passes, achieved with a fixed probe energy and without increasingly informative probe states. The authors support the theory with an optical experiment measuring photon rotation, where Q-plates and a Dove prism realize the indefinite-time-direction generating process and the parameterizing rotation.","feed_headline":"Metrology precision scales as 1/T² via time-direction superposition","feed_subtitle":"A quantum switch before the sensing step turns noncommuting operations into quadratic precision gain, shown on photon rotation.","key_machinery":"The quantum switch: a two-level ancilla in a superposition of |0⟩ and |1⟩ routes the probe through U_C or U_C†, creating a coherent superposition of forward and backward time directions during the generating process. The relevant identity is the shifted canonical momentum K̄_I = (K̄_S ∓ T_C T_S) on the two ancilla branches, which follows from [H_C, V_S] = i and makes the momentum uncertainty grow as T⁴. This is what converts a linearly growing resource (time) into a quadratically growing precision.","core_discovery":"The central claim is that an ancilla-mediated generating process with indefinite time direction, placed before the parameterizing unitary, increases the canonical-momentum uncertainty ΔK̄ nonlinearly. Concretely, for a parameterizing Hamiltonian H_S with characteristic operator V_S = ∂_g H_S and a generating Hamiltonian H_C satisfying [H_C, V_S] = i, the canonical momentum in the Heisenberg picture becomes K̄_I = (K̄_S − T_C T_S) ⊗ |0⟩⟨0| + (K̄_S + T_C T_S) ⊗ |1⟩⟨1|, so the variance of K̄_I acquires a T_C² T_S² term. Setting T_C = T_S = T gives δg ≥ 1/(2√ν √(ΔV_S² T² + T⁴)), i.e., δg ∝ 1/T² when the T⁴ term dominates; repeating the joint process N times gives δg ∝ 1/N². The experiment with p","pith_inferences":["Editorial extension: the 'resource-free' claim is sensitive to the choice of resource measure. The paper bounds the decomposition-averaged uncertainty ΔV̄_S of the characteristic operator on the reduced probe state, whereas the standard resource count in the Heisenberg limit is the variance of the parameter generator on the full sensing state; on that measure, the same calculation shows ΔK̄_I incl","Editorial extension: because the final probe-ancilla state's Fisher information is F(g)=4T⁴ (for the OAM experiment), the protocol can be interpreted as a standard quantum channel whose QFI is quadratic in time; equivalently, the same precision could be obtained by choosing a probe with generator variance T⁴, so the practical gain is in realizing that variance with a fixed-energy ancilla rather th","Editorial extension: the framework suggests testing the same mechanism with other conjugate pairs, e.g., photon-number and phase, where the generating process would shift photon number and the parameterizing process would be a phase shift; the same [H_C,V_S]=i algebra predicts the same 1/T² scaling.","Editorial extension: an experimental falsifier would be a direct tomographic estimate of the QFI of the final state as a function of T; the paper's probabilities already imply F=4T⁴, so measuring F(T)≠4T⁴ would indicate a departure."],"forward_implications":["If the central claim is correct, quantum metrology can beat the customary 1/T and 1/N scalings using time-independent Hamiltonians, contrary to the standard belief.","The scheme is experimentally accessible with photonic OAM and polarization, needing only Q-plates, mirrors, and a Dove prism, with no energetic probe excitation.","Sequential application of the joint process gives a quadratic improvement in N, so repeated queries of the same gate can be exploited more efficiently than the linear bound.","The uncertainty-principle formulation unifies previously reported super-Heisenberg regimes (nonlinear interactions, time-dependent control, indefinite causal order) as cases where ΔK̄ grows nonlinearly.","The bound is consistent with and can be attained within the quantum Cramér–Rao bound framework, with equality when the initial state diagonalizes K̄."],"fun_headline_variants":["Quantum switch boosts metrology precision to 1/T^2 scaling","Time-direction superposition yields quadratic metrology gain","Indefinite-time encoding gives 1/T^2 metrology without probe resources","Quadratic precision scaling from time-direction superposition","No probe resources needed: time switch gives 1/T^2 metrology"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The bound relies on measuring resources by the average uncertainty of the characteristic operator V_S on the reduced probe state; if the variance of the parameter generator on the full probe–ancilla state is the resource to be counted, the advertised 'resource-free' nonlinear advantage disappears because that variance already grows as T⁴.","fun_headline_variants_meta":{"raw":{"variants":["Quantum switch boosts metrology precision to 1/T^2 scaling","Time-direction superposition yields quadratic metrology gain","Indefinite-time encoding gives 1/T^2 metrology without probe resources","Quadratic precision scaling from time-direction superposition","No probe resources needed: time switch gives 1/T^2 metrology"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000774,"raw_usage":{"total_tokens":3289,"prompt_tokens":795,"completion_tokens":2494,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":539,"completion_tokens_details":{"reasoning_tokens":2408}},"tokens_in":539,"tokens_out":2494,"duration_ms":14352,"temperature":1.0,"reasoning_tokens":2408,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T10:31:24.426525+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For the OAM experiment, perform full tomography of the probe–ancilla state at T=1,2,3,4 and compute the quantum Fisher information F(g) of the parameter. The paper's own probabilities give F=4T⁴; if a measurement shows F growing as T⁴, the enhancement is carried by the generator variance on the sensing state, and the 'resource-free' status depends entirely on the paper's chosen measure on V_S.","supporting_citations":[],"review_version":1}