{"id":"a18e5812-2499-4f41-b91b-fe3bf594b9d6","arxiv_id":"2601.14708","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A single optical probe sent in a probabilistic mixture of two opposite directions around a cyclic sensor network can estimate the average tilt with 1/N² precision scaling, without entangled probes.","lead":"A proposed sensing network sends one light beam around a loop of sensors in two opposite directions and claims this makes the precision scale as 1/N² with the number of sensors. The authors report free-space beam-tilt measurements on up to nine sensors at picoradian-level sensitivity.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Experimental scaling is confounded: synchronized identical tilts make the fixed-order baseline also N², so the switch's advantage for independent parameters is not demonstrated.","rationale":"The reader's primary weakest_assumption — that the QCRB is computed at g1=g2=0 and may fail for finite tilts — does not land for the full joint-state QFIM. The generators H1(g), H2(g) differ from their g=0 values only by c-number shifts (from the exact Heisenberg-algebra product structure), so the covariance matrix, and hence the QFIM, is independent of g. The small-parameter caveat is limited to the probe-alone analysis of Supplementary Note 1.5. However, the reader's secondary remarks — no independent baseline and no non-uniform sensor parameters — point toward a more serious issue: the experiment uses synchronized identical tilts, which turns the problem into single-parameter estimation. In that regime, a fixed-order sequential protocol already achieves 1/N² precision via the same noncommutative propagation amplification, so the observed N² scaling cannot be attributed to the causal-order switch. The theory for estimating an average of independent parameters remains internally consistent, so the verdict should remain CONDITIONAL, but the condition should be reframed: the experimental evidence must compare against the correct fixed-order baseline for common tilts, or test non-uniform parameters. Since the reader's verdict was already CONDITIONAL, I recommend no change to the verdict label, but the rationale should be updated to this more load-bearing concern.","tokens_in":26428,"tokens_out":26846,"duration_ms":287717,"concrete_test":"Compute the QFI for estimating a common tilt φ (θ_j = φ for all j) from the fixed-order evolution U+ of Eq. (S12) using the exact Heisenberg-algebra expression, and compare it with the switch evolution Eq. (S19). If both yield QFI ~ N^4, the switch provides no advantage for the common-tilt task. Experimentally, add a control configuration that blocks the reverse path (removing the switch), apply the same synchronized 10 kHz tilts, and measure SNR vs N for N=1..9; fit to Eq. (15). If the fixed-path SNR also scales as N², the central experimental claim that the switch surpasses the Heisenberg limit is confounded.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The experimental section applies synchronized identical tilts to all sensors, reducing the task to estimating a single common parameter φ. For that task, the fixed-order sequential protocol already produces a propagation-induced spatial shift proportional to N² (the same noncommutativity effect), with QFI scaling as N^4 and precision as 1/N². The paper's '1/N Heisenberg' baseline is derived from the two-parameter QFIM for estimating the average of N independent unknown θ_j (Eqs. S15–S18), which is not the relevant baseline for the common-tilt experiment. Thus the observed N² SNR scaling in Figs. 4–6 does not demonstrate an advantage of causal-order switching over a fixed causal order; it merely demonstrates the geometric amplification that a single-direction pass would also exhibit. The switch may still enhance precision for independent sensor parameters, but that scenario is not tested. This is a mismatch between the theoretical quantity (average of independent parameters) and the experimental quantity (common synchronized tilt).","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a distributed quantum sensing protocol in a cyclic network where a single probe sequentially interrogates N sensors in two opposite causal orders, controlled either by a quantum or a classical switch ancilla. The central theoretical claim is that noncommutativity between free-space propagation and local momentum-kick sensing turns propagation into a metrological resource, yielding a 1/N^2 precision scaling for the average of independent parameters without multipartite entanglement. The authors report an experimental implementation using a free-space Sagnac interferometer with up to 9 sensors and synchronized beam tilts, claiming picoradian-level precision and a fitted scaling δφ_min ≈ 4.77/(N^2+4.25N) nrad. The supplementary materials contain the QFIM calculations for the fixed-order, quantum-switch, and classical-switch protocols, as well as a small-parameter analysis of estimating from the probe alone.","tokens_in":26690,"tokens_out":7225,"duration_ms":82456,"significance":"If the theoretical claim is correct, this is a substantial conceptual advance: it identifies propagation noncommutativity and causal-order mixing as resources that can beat the conventional 1/N Heisenberg scaling for distributed sensing without entangled probes. The supplementary QFIM derivations are internally coherent and provide explicit scaling constants, and the classical-switch result is a nontrivial extension of prior indefinite-causal-order metrology. The experimental effort, with reconfigurable up-to-9-sensor network and weak-value readout, is also potentially valuable. However, as detailed below, the experimental validation is mismatched with the theoretical claim: the experiment applies synchronized identical tilts, which reduces the task to estimating a single common parameter, for which the fixed-order protocol already exhibits the same N^2 geometric amplification. Thus the experimental demonstration, as presented, does not support the central advantage of causal-order switching.","major_comments":[{"comment":"The experiment applies synchronized identical tilts to all sensors, so it estimates a single common parameter φ. For this task, the fixed-order evolution in Eq. (S12) already produces a momentum displacement g1/k with g1 = φ Σ_{j=0}^{N-1} z_j (N−j), which scales as O(N^2) for equal spacings. Thus a fixed-order sequential network would also produce an SNR that grows ~N^2, and the observed scaling in Figs. 4–6 does not demonstrate an advantage of causal-order switching. The theoretical fixed-order QCRB in Eqs. (S17)–(S18) applies to estimating the average of N independent unknown θ_j, not to the synchronized-common-tilt experiment. The manuscript must either test independent parameters or compare the switched network against a fixed-order baseline under the same common-tilt conditions.","section":"Section II, Eq. (14); Supplementary Eq. (S12)"},{"comment":"The '1/N–scaling precision limit' shown in Fig. 6 is generated by replacing the N^2 term with 1 in the same fitted relation δφ_min ≈ 4.77/(N^2+4.25N) that is the paper's own claim. This is not an independent baseline: it uses the fitted coefficients from the switched-network data and therefore cannot serve as a comparison. A valid Heisenberg-limit baseline must be derived and preferably measured for the fixed-order protocol under identical conditions (same WVA, same detection, same common-tilt signal). Without this, the statement that the experiment surpasses Heisenberg scaling is not supported.","section":"Fig. 6; Section II 'Experimental results'"},{"comment":"The claim that the 1/N^2 scaling 'still holds' for estimating the average from the final mixed probe state alone is qualified in Supplementary Note 1.5 as being derived only at g1=g2=0: 'it is equivalent to calculate the QFIM at g=0.' The main text and abstract do not prominently state this small-parameter restriction. Since this caveat is explicitly acknowledged in the supplement, it should be brought into the main text to avoid overstating the classical-mixture result. The full joint-state classical-switch QCRB in Eq. (S29) appears exact, so this is a limitation of the trace-out version rather than of the central joint-state result.","section":"Discussion; Supplementary Note 1.5"}],"minor_comments":[{"comment":"The spectral decomposition states '|Ψ−⟩ = |ψ−⟩ ⊗ |0⟩'; this should be |ψ−⟩ ⊗ |1⟩. The subsequent orthogonality argument relies on the ancilla states being orthogonal.","section":"Supplementary Note 1.4"},{"comment":"The axis labels are garbled ('103 (peJd) U!'); they should be readable and include units.","section":"Fig. 6"},{"comment":"The noise-floor independence assumption is stated as 'approximately independent of the number of sensors' but is not directly characterized. With increasing N the interferometer path length and number of mirrors change; even with stabilized received power, technical noise could vary. A direct measurement of the noise floor versus N would strengthen the experimental scaling claim.","section":"Methods B"},{"comment":"There are several typographical errors: 'sinuous' should be 'sinusoidal', 'spec trums' should be 'spectrums', 'of of' appears in Fig. 4 caption, and some equation references in the main text are to the supplement without equation numbers.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The theoretical core of the paper is interesting and the QFIM calculations appear sound, but the experimental section currently validates a different statement than the one emphasized in the abstract and title. The circular baseline in Fig. 6 and the common-tilt versus independent-parameters mismatch are load-bearing issues. I would be willing to reconsider after the authors either provide a fixed-order experimental baseline under identical conditions or reframe the experimental contribution to the common-parameter geometric amplification, while clearly distinguishing it from the independent-parameter advantage."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The theoretical claim is genuinely new: a probabilistic mixture of opposite causal orders (a classical switch) can match the 1/N² precision scaling of a coherent quantum switch for estimating the average of N independent distributed parameters, without entanglement. The QFIM derivations in the supplement are internally consistent: the fixed-order baseline is 1/N, and the classical-switch QCRB asymptotically goes as 1/N² from the same noncommutativity between propagation and momentum kicks. That part is worth taking seriously.\n\nThe experimental section is the soft spot. The setup applies synchronized identical tilts to all sensors, so the task is estimating a single common parameter. For that task, the fixed-order sequential protocol already produces the same N² spatial shift — the same noncommutativity effect — so the observed SNR scaling with N does not demonstrate any advantage from causal-order switching. The '1/N Heisenberg baseline' in Fig. 6 is constructed by replacing the N² term with 1 in the fitted formula, not by an independent calculation or measurement for the actual task. That comparison is circular. The theory for independent parameters is simply not tested.\n\nTwo smaller issues: the classical-switch QCRB is evaluated at g=0 and assumed to hold for small finite parameters — the supplement states this, but the experimental tilts are finite, so the scaling at realistic signal levels remains unverified. And the claim that the noise floor is independent of N is plausible but not directly demonstrated.\n\nNet: the classical-switch result is a real theoretical contribution, and the paper is coherent on its own terms. But the abstract's claim of experimentally surpassing the Heisenberg limit overreaches. I would send this to peer review, because the theory deserves a careful referee and the experimental claim needs to be either reanalyzed with a proper fixed-order baseline for common tilts or replaced by a test with independent parameters. A firm referee could turn this into a solid paper by separating the theoretical result from the overclaimed experimental demonstration.","headline":"Solid theory on classical-switch distributed sensing; the experiment tests only common synchronized tilts, so the claimed demonstration of switch advantage does not hold as presented.","tokens_in":27131,"tokens_out":3324,"would_cite":true,"duration_ms":40214,"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":"A probe that loops a sensor network in both causal directions can estimate the average parameter with 1/N² precision, beating the 1/N Heisenberg scaling, without using entangled sensors.","keywords":["distributed quantum sensing","causal order","super-Heisenberg scaling","quantum Fisher information","beam-tilt sensing","noncommutativity","cyclic network","weak value amplification"],"falsifier":"Measure the estimation precision (or the quantum Fisher information) for a fixed number of sensors N as the applied tilt angle θ̄ is increased from near zero to larger values. If the precision degrades faster than the prediction based on the g=0 bound, the claimed 1/N² scaling does not hold for finite signals.","tokens_in":26361,"feed_emoji":"🔬","tokens_out":2951,"duration_ms":34256,"temperature":0.7,"pith_summary":"The paper introduces a distributed sensing protocol in which a single probe travels around a cyclic network of N sensors in both forward and backward directions, controlled by a switch ancilla. Because free-space propagation and the local momentum-kick sensing operation do not commute, the round-trip routing converts propagation from a passive loss into an active metrological resource that amplifies the accumulated signal by an N² factor. The authors claim the quantum Cramér–Rao bound for estimating the average tilt angle scales as 1/N², both for a coherent quantum switch and for a simple probabilistic mixture of causal orders. This super-Heisenberg scaling does not require multipartite entanglement, and the paper demonstrates it experimentally in a free-space optical network of up to 9 beam-tilt sensors, reaching picoradian-level precision. If correct, this would establish routing geometry and propagation dynamics as practical, scalable resources for distributed quantum sensing.","feed_headline":"Two-way loop sensing hits 1/N² precision without entanglement","feed_subtitle":"A single probe, sent around a cyclic network in both orders, turns propagation into a measuring resource.","key_machinery":"The central object is the noncommutativity between the sensing Hamiltonian (momentum kick, exp(−iθX)) and the free-space propagation Hamiltonian (exp(−izP²/2k)). This noncommutativity converts the propagation distance into an effective momentum amplification: the probe acquires a spatial displacement proportional to (z̄/2k)N²θ̄ when traversed in both causal orders. A switch ancilla—initialized either as a coherent superposition or as a classical mixture—controls the direction of traversal; a parity operation on the reverse branch plus weak-value post-selection (with imaginary weak value Aw ≈ i/ε) provide the experimental readout, mapping the amplified spatial shift to a measurable transverse","core_discovery":"The central claim is that by sending the probe through a cyclic network in a superposition—or even in a classical probabilistic mixture—of two opposite causal orders, the estimation precision for the average parameter θ̄ improves as δθ̄ ∝ 1/N². The enhancement arises because the momentum-kick sensing unitaries exp(−iθX) and the free-space propagation unitaries exp(−izP²/2k) do not commute, so the propagation displaces the probe by an amount proportional to the accumulated momentum kick, and traversing the network in both orders adds these displacements constructively. The asymptotic QCRB is shown to be lim_{N→∞} C_{θ̄}/N⁻⁴ = k²/(z̄² ⟨ΔP²⟩ᵢ), independent of whether the switch is quantum or cl","pith_inferences":["If the zero-parameter QFI bound holds for finite tilts, similar noncommutativity between propagation and non-commuting sensing operations (e.g., dispersion and phase shifts in fibers) could yield analogous N² enhancements in other continuous-variable sensing platforms.","The paper's demonstration that a classical mixture of orders matches the quantum switch suggests that the phenomenon is more about the geometry of opposing propagation paths than about indefinite causal order; this may simplify practical deployment but also invites scrutiny about whether the resource is truly quantum.","The scaling relies on the noise floor being independent of N; a natural test is to measure the noise floor as sensors are added. If technical noise grows with N, the practical advantage will saturate earlier than the zero-noise prediction.","The weak-value amplification used here converts the amplified spatial shift into an imaginary weak value; the same technique could be used to probe other nonlinear signal accumulations in optical networks, potentially extending the approach to distributed phase or displacement sensing."],"forward_implications":["Distributed quantum sensing can achieve super-Heisenberg scaling without multipartite entanglement, using a single coherent probe and a cyclic network.","A classical probabilistic mixture of causal orders suffices for the 1/N² scaling; no coherent quantum switch is required for the theoretical limit.","The protocol is robust to classical noise and can be implemented with coherent light sources, avoiding the fragility of squeezed or entangled states.","The demonstrated picoradian-level beam-tilt sensitivity in a 9-sensor network suggests practical applications in interferometric alignment, vibration monitoring, and acoustic sensing.","The prediction that measured SNR grows quadratically with sensor count provides a clear, directly testable signature of the enhancement."],"fun_headline_variants":["No entanglement needed: 1/N² precision via two-way routing","Quantum sensing gets quadratic boost from bidirectional loops","Single probe, two directions: precision scales as N²","Routing beats entanglement for distributed quantum sensing"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The 1/N² quantum Cramér–Rao bound is derived at the single point g₁=g₂=0 (zero tilt), and the paper assumes this also represents small but finite tilt amplitudes; if the quantum Fisher information drops away from g=0, the asymptotic scaling would not describe realistic signals.","fun_headline_variants_meta":{"raw":{"variants":["No entanglement needed: 1/N² precision via two-way routing","Quantum sensing gets quadratic boost from bidirectional loops","Single probe, two directions: precision scales as N²","Routing beats entanglement for distributed quantum sensing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000512,"raw_usage":{"total_tokens":2310,"prompt_tokens":712,"completion_tokens":1598,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":456,"completion_tokens_details":{"reasoning_tokens":1536}},"tokens_in":456,"tokens_out":1598,"duration_ms":16269,"temperature":1.0,"reasoning_tokens":1536,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T09:07:51.614042+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the estimation precision (or the quantum Fisher information) for a fixed number of sensors N as the applied tilt angle θ̄ is increased from near zero to larger values. If the precision degrades faster than the prediction based on the g=0 bound, the claimed 1/N² scaling does not hold for finite signals.","supporting_citations":[],"review_version":1}