{"id":"c1b16b5d-87e7-491b-a662-73d4f5e69198","arxiv_id":"2607.26502","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Noise that changes the sensing-basis label can be converted to erasures and removed by a passive ancilla projection, while label-preserving noise cannot.","lead":"This paper shows that certain kinds of in-place quantum sensing noise can be converted into detectable erasures with only a passive ancilla, and demonstrates the idea in a photonic phase-sensing experiment. The result matters because it offers a low-overhead way to recover precision lost to realistic noise.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Necessary direction of Theorem I is asserted, not proved: the paper never defines the allowed class of passive schemes or the precise sense in which diagonal noise is indistinguishable from signal; the boundary may depend on whether the retained subspace is rank-d and whether erasure outputs are dis","rationale":"The reader's weakest assumption is exactly where the paper is least secure. The sufficient construction (V1,V2,M_s) is explicit and correct: off-diagonal operators break the system-ancilla label match, so Eq. (5) holds. The necessary half, however, is argued heuristically. The text in 'Non-convertible components' asserts that a diagonal term 'produces the same type of output state as the signal evolution' and that 'there is no extra information' to distinguish it, but this is not derived from a definition of the allowed operations. The Discussion's admission that discarded erasure outputs may carry extractable information (and that recycling them is open) makes clear that the boundary is for discard-only passive schemes; the theorem should state this qualification. I checked whether a formal no-go is available: if a passive scheme is defined by a fixed encoding isometry V1, fixed decoding V2, and a rank-d retained projector M_s = I_S⊗|0><0|_A, then signal preservation for all θ forces C = V1(H_S⊗|0>) to be contained in the retained subspace; equal dimensions give C = C_ret, and U_θC⊂C implies C is G-invariant. Any D∈D_G then preserves C, so no diagonal term can be projected to the erasure subspace. Thus the claimed boundary is plausible and probably correct for that class. The concern is therefore not a counterexample but a missing formal proof and a missing precise statement of the class. This is exactly what a conditional acceptance should require. The experimental demonstration's use of a QWP rather than a true stochastic Pauli channel is a secondary issue; the theoretical no-go is the load-bearing point.","tokens_in":13387,"tokens_out":25916,"duration_ms":209227,"concrete_test":"Formalize a passive scheme as (V1,V2,M_s) with M_s = I_S ⊗ |0><0|_A and prove: if V2(U_theta⊗I)V1(H_S⊗|0>) ⊂ H_S⊗|0> for all θ, then C = V1(H_S⊗|0>) is G-invariant; hence for any D,D'∈D_G, Tr[M_s V2(D⊗I)ρ_enc(D'^dagger⊗I)V2^dagger M_s] ≠ 0 for some ρ, so no diagonal term is erasable. If the proof fails, construct a counterexample (e.g., with a 4-dimensional retained subspace or a θ-dependent final measurement); existence of such a counterexample would narrow Theorem I to the specific V1,V2,M_s construction rather than all passive schemes.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing issue is the 'only if' half of Theorem I. The proof in 'Non-convertible components of in-space noise' argues that a term E_mu(.)E_nu^dagger with both E_mu,E_nu in D_G just multiplies each matrix element by d_{mu,x} d*_{nu,x'}, 'the same type of output state as the signal evolution,' and that therefore no erasure flag can be constructed. This is an intuition, not a theorem: a clever encoding might in principle make the ancilla sensitive to the phase pattern d_{mu,x} d*_{nu,x'} without disturbing U_theta, and the paper gives no formal definition of 'passive scheme' or of 'without damaging the signal' under which such schemes are excluded. A natural formalization is a fixed encoding isometry V1, a fixed decoding V2, and a single rank-d retained projector M_s = I_S ⊗ |0><0|_A. Under that definition the no-go can be proven: U_theta-invariance of the retained subspace forces V1(H_S⊗|0>) to be G-invariant, and every diagonal D∈D_G preserves that subspace, so D(.)D'^dagger terms cannot be orthogonal to the retained outcome. But if one allows a larger retained subspace or a θ-dependent final measurement, the argument collapses, and the paper's Discussion even acknowledges that recycled erasure outputs may carry information. Thus the iff statement is broader than what is actually demonstrated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper addresses the conversion of in-space quantum sensing noise into erasures. It re-expresses a general noise channel in an operator basis adapted to the sensing generator G, decomposing the operator space into the diagonal subspace D_G and the off-diagonal subspace O_G. The central claim, Theorem I, is that under a passive scheme a noise-channel term E_μ(·)E_ν† is erasure-convertible if and only if at least one of E_μ, E_ν lies in O_G. The authors construct an explicit ancilla-based encoding/decoding scheme that maps all 'convertible' terms to an orthogonal ancilla subspace, which is discarded by a final projection. They analyze the effective QFI for phase-damping and amplitude-damping channels and report a photonic proof-of-principle experiment using polarization as the sensing qubit and orbital angular momentum as the ancilla, claiming restoration of standard-quantum-limit precision for an emulated Pauli-noise channel of convertible weight 0.5.","tokens_in":13837,"tokens_out":6460,"duration_ms":103464,"significance":"If Theorem I can be made rigorous, the result would provide a clean, noise-model-agnostic boundary for passive erasure conversion in quantum sensing: off-diagonal (label-changing) terms can be converted to erasures, while diagonal terms cannot. The sufficient construction is explicit, deterministic, and requires no active control or noise-model knowledge, and the QFI formulas follow from stated channel assumptions with no fitted parameters. The experiment, although using a deterministic emulation, demonstrates the optical routing and filtering mechanism with data that track the parameter-free theoretical curve. These are real strengths. The main weakness is that the necessity half of the central theorem is asserted through an informal indistinguishability argument rather than proved under a precise definition of the allowed operations.","major_comments":[{"comment":"The 'only if' direction of Theorem I is not proved. The text argues that a diagonal term E_μ(·)E_ν† only multiplies density-matrix elements and 'there is no extra information that can distinguish this noise contribution from the signal evolution,' but no formal definition is given of the allowed class of passive schemes or of 'without damaging the signal.' A clever fixed encoding/decoding with a larger retained subspace, or a θ-dependent readout, might in principle detect the pattern d_{μ,x}d*_{ν,x'} without disturbing U_θ; the paper neither proves such schemes impossible nor states the restriction under which they are excluded. This is load-bearing because the theorem's 'iff' is the paper's central claim. A formal definition (e.g., fixed V_1, V_2, and final projection M_s = I_S⊗|0><0|_A, with erasure outputs discarded) would make the necessity provable; please provide it.","section":"Non-convertible components of in-space noise"},{"comment":"Eqs. (15)-(16): the experiment uses a deterministic unitary QWP, not a stochastic Pauli-noise channel. Eq. (16) contains coherent cross terms i/2(ρσ_n - σ_nρ); the equivalence to F_{1/2}(ρ)=1/2ρ+1/2σ_nρσ_n holds only after the final ancilla projection and only for the retained measurement statistics. The abstract's claim of demonstrating 'a Pauli-noise channel with erasure-convertible weight 0.5' therefore overstates the demonstration. The experiment validates the optical routing/filtering mechanism for a unitary rotation, but does not test conversion of incoherent stochastic Pauli noise. Please either implement a genuinely mixed channel or clearly qualify the claim as a proof-of-principle under deterministic emulation.","section":"Methods: QWP emulation of a channel with fixed erasure-convertible weight"}],"minor_comments":[{"comment":"Please state explicitly that p_s(θ) is the retained-outcome probability under the full noisy channel and that ρ_θ^{(s)} is the normalized retained state. The current notation leaves this implicit.","section":"Theoretical QFI analysis, Eq. (6)"},{"comment":"The 'total coefficient weight' W_conv(R)=Σ' |χ^{(R)}_{μν}|² is not obviously the correct physical measure of the convertible fraction, especially for cross terms that do not correspond to independent outcomes. Please justify this choice or define it operationally.","section":"Methods: Optional optimization of G"},{"comment":"Several central derivations and robustness claims are deferred to the Supplementary Material (derivation of Eq. (7), ancilla-noise tolerance, GHZ probes, far-field propagation). The supplementary material was not part of the submitted manuscript text; please include it or provide proof sketches in the main text.","section":"General"},{"comment":"Please specify how the error bars and the 220-repetition grouping into five batches were converted into the reported precision; the statistical procedure is currently described only in the Supplementary.","section":"Experimental results, Fig. 8"}],"recommendation":"major_revision","confidential_remarks":"The paper has a valuable sufficient construction and a plausible necessary condition, but the central theorem's necessity half needs a rigorous formalization before the 'iff' claim can be accepted. The experimental overstatement about a 'Pauli-noise channel' should also be corrected. I see no indication of bad faith or circularity; the issues are technical and fixable within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read. The paper gives a clean, generator-dependent division of in-space noise into erasure-convertible and non-convertible terms, plus a static ancilla-flag scheme that provably discards the convertible terms. The sufficient direction is proven cleanly, the QFI analysis for phase and amplitude damping is consistent, and the experiment—using OAM as the ancilla—is a credible proof-of-principle showing SQL-level precision recovery. No free parameters are fitted; the QFI formulas follow from the stated assumptions. The reader's soundness score of 4 is a bit harsh; the construction is sound.\n\nThe real soft spot is the necessary half of Theorem I. The paper argues that diagonal terms produce the same type of output state as the signal and therefore cannot be flagged. That is an intuition, not a proof. There is no formal definition of the allowed class of passive schemes, and no no-go theorem ruling out larger retained subspaces or θ-dependent readouts. The stress-test note is right that under the natural formalization—fixed encoding, fixed decoding, rank-d retained subspace—the no-go goes through. So I think the theorem is likely correct for the intended setting, but the stated 'if and only if' is broader than what is demonstrated. This is a moderate issue: fixable by formalizing the class of operations or softening the claim.\n\nThe experimental emulation is a deterministic QWP unitary, not a real stochastic Pauli channel. The paper explains the equivalence after the ancilla projection, and for a proof-of-principle that is acceptable, but it is a caveat.\n\nWho's it for? People working on noise-resilient quantum sensing, especially erasure-based approaches. Worth a serious referee; I'd send it out with a request to formalize the necessary direction or temper the theorem, and to clarify the experimental emulation's scope. I'd cite the criterion and would bring the paper to reading group.","headline":"A clean, generator-dependent criterion for converting in-space sensing noise to erasures, with a simple passive construction and a convincing proof-of-principle experiment; the main theorem's necessary direction is argued rather than proven, so the strongest claim outruns the evidence by a step.","tokens_in":14215,"tokens_out":7505,"would_cite":true,"duration_ms":139165,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P45","81P50"],"pacs":["03.67.-a"],"model":"deepseek-v4-flash","headline":"Theorem I: a noise-channel term is erasure-convertible if and only if it flips a sensing-basis label on at least one side; both-diagonal terms are inseparable from the signal.","keywords":["erasure conversion","quantum sensing","in-space noise","quantum Fisher information","passive noise mitigation","orbital angular momentum","phase estimation","standard quantum limit"],"falsifier":"A concrete counterexample would be a passive ancilla strategy that successfully flags and discards a channel term such as σ_z ρ σ_z (both operators diagonal in the σ_z eigenbasis) without degrading the signal, contradicting the paper's non-convertible sector. Experimentally, one could engineer a channel containing only diagonal noise terms and check whether any passive protocol, including the scheme here, can raise the QFI above the unmodified noisy channel; the paper predicts no improvement is possible.","tokens_in":13365,"feed_emoji":"⚛️","tokens_out":7752,"duration_ms":58850,"temperature":0.7,"pith_summary":"This paper asks a sharp question: which parts of a noise channel that acts entirely inside the sensing Hilbert space can be converted into erasures — errors whose occurrence is flagged — and removed without destroying the signal? It answers with a necessary and sufficient condition tied to the sensing generator: a channel term is erasure-convertible if and only if at least one of its two operators changes the eigenbasis label of the generator; if both operators are label-preserving, the noise is structurally identical to the signal and cannot be pulled out passively. The authors construct a passive scheme — a fixed ancilla encoding and a single final projection — that realizes the conversion for all convertible terms, requiring no noise-model knowledge, active control, or precise timing. The theory is confirmed in a single-photon phase-sensing experiment using orbital angular momentum as the ancilla, where the recovered precision matches the theoretical curve and approaches the standard quantum limit despite a Pauli-noise channel with erasure-convertible weight 0.5.","feed_headline":"Passive scheme turns sensing noise into discardable erasures","feed_subtitle":"No active control or noise model needed: experiment recovers the precision limit.","key_machinery":"The carrying object is the Hilbert-Schmidt decomposition of the system operator space L(H_S) into D_G = span{|x⟩⟨x|} and O_G = span{|x⟩⟨y|, x≠y}, defined by the eigenbasis of the sensing generator G. Every in-space noise channel is re-expressed in these bases as N(·) = Σ χ_{μν} E_μ(·) E_ν†, and the erasure-conversion condition classifies each channel term by whether E_μ or E_ν lies in O_G. The implementing mechanism is a static ancilla encoding V1 = Σ_x |x⟩⟨x| ⊗ U_{κ(x)} that records the sensing-basis label in orthogonal ancilla states, together with the decoding V2 and the projection M_s = I ⊗ |0⟩⟨0|: any term that changes a label maps outside the retained ancilla mode, so a single projecti","core_discovery":"The central claim is Theorem I: under a passive scheme, a noise-channel component E_mu(·)E_nu† can be removed from the retained sensing outcome as an erasure without damaging the signal if and only if E_mu or E_nu belongs to the off-diagonal subspace O_G, i.e., at least one side mixes the eigenbasis labels of the sensing generator G. Terms with both operators in the diagonal subspace D_G act on each matrix element by multiplying its complex coefficient — exactly the same type of transformation produced by the signal evolution U_theta — so no passive readout can flag them without also discarding the signal. The sufficient direction is proven by explicit construction: an encoding V1 writes the","pith_inferences":["If the 'only if' direction is made into a formal no-go theorem, the diagonal/off-diagonal split would define an information-theoretic boundary for all non-adaptive, non-feedback sensing schemes, and could be used to certify when active control is genuinely necessary.","The same ancilla-index encoding may be adapted to multi-parameter estimation by choosing a full basis that diagonalizes a set of commuting generators; for non-commuting generators, a tradeoff between erasure-conversion and simultaneous estimation seems likely.","A possible extension is to recycle the discarded erasure outputs: because the final projection destroys information, a hybrid scheme that measures the erased part and feeds it back could in principle beat the passive bound, which the paper leaves open."],"forward_implications":["Erasure-convertible weight of a noise channel is well-defined and can be maximized by reorienting the sensing generator among physically equivalent bases, without changing the noiseless QFI.","When all noise lies in T_conv, the effective QFI equals the noiseless QFI times the retained probability; for phase damping at p=0.5, this recovers half the ideal QFI whereas the raw channel carries none.","The final ancilla projection discards all convertible and cross terms; residual precision is set strictly by the non-convertible weight, so the protocol is useful whenever that weight is low.","In the photonic demonstration, precision follows the theoretical curve and approaches the SQL, confirming that the condition is not just abstract but experimentally realizable with OAM as a passive ancilla."],"fun_headline_variants":["Passive erasure conversion restores quantum sensing precision","Noise becomes erasures: passive scheme hits precision limit","Convert sensing noise to erasures without active control","Erasure conversion: passive route to robust quantum sensing","Turn noise into erasures, recover standard quantum limit"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The 'only if' direction assumes that no passive scheme can extract any extra information — from environment monitoring, feedback, or intermediate measurements — to distinguish label-preserving diagonal noise from the signal at the final readout; if such information were available within the passive class, the claimed necessary condition would fail.","fun_headline_variants_meta":{"raw":{"variants":["Passive erasure conversion restores quantum sensing precision","Noise becomes erasures: passive scheme hits precision limit","Convert sensing noise to erasures without active control","Erasure conversion: passive route to robust quantum sensing","Turn noise into erasures, recover standard quantum limit"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000165,"raw_usage":{"total_tokens":1060,"prompt_tokens":690,"completion_tokens":370,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":434,"completion_tokens_details":{"reasoning_tokens":293}},"tokens_in":434,"tokens_out":370,"duration_ms":11616,"temperature":1.0,"reasoning_tokens":293,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T14:19:50.646551+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete counterexample would be a passive ancilla strategy that successfully flags and discards a channel term such as σ_z ρ σ_z (both operators diagonal in the σ_z eigenbasis) without degrading the signal, contradicting the paper's non-convertible sector. Experimentally, one could engineer a channel containing only diagonal noise terms and check whether any passive protocol, including the scheme here, can raise the QFI above the unmodified noisy channel; the paper predicts no improvement is possible.","supporting_citations":[],"review_version":1}