{"id":"5c49504d-456f-4eeb-b9a4-a3b390d31e34","arxiv_id":"2608.01114","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A conic and semidefinite-programming framework computes exact Holevo, Nagaoka-Hayashi, and SLD precision bounds for multiparameter quantum metrology across parallel, sequential, and indefinite-causal-order strategies, with energy and memory constraints.","lead":"This paper develops a computational framework that finds the best way to estimate several unknown quantities at once in quantum sensors, by jointly optimizing the probe state, the control operations, and the final measurement. It computes fundamental precision limits for different sensing strategies, including exotic ones with indefinite causal order, and demonstrates which strategy class wins in a noisy magnetometry example.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Appendix D2 regularization is internally inconsistent: the added identity term in Eq. (D23) is normalized by ∏_{k=1}^{2N} d_k, but Eq. (D24) uses ∏_{k=1}^{N} d_k, so the constructed P_ε violates the sequential comb constraints (A1); the proof of SP_k = QP_k is incomplete as written.","rationale":"The reader's weakest_assumption targeted exactly Appendix D's cone-preservation and purification step; my check isolates a concrete failure in that step as printed. The central claim (SP_k = QP_k for k=2,3,4) is the foundation for all exact-SDP statements, so a gap here is load-bearing. The good news is that the noiseless analytical benchmarks (Fig. 2(c), Appendix F) support the formulation in the full-rank regime, and the regularization identity can likely be repaired by using the product of input-space dimensions, under which I_{S̄}/∏ d_{2k−1} satisfies the recursive comb conditions. This is why I do not move the verdict to REJECT: the flaw is specific and plausibly a typo, but it must be fixed and independently checked before the exactness claim can be accepted. The paper also ships no code or machine-checked proof, so the proof-sketch status of Appendix D is the main source of correctness risk. The abstract's unqualified 'strict hierarchy' and the ICO subclass formulas (which appear to be convex mixtures rather than coherent superpositions) are additional concerns, but they are secondary to the central SDP equivalence. Overall, the reader's CONDITIONAL verdict stands; my analysis makes the condition concrete: correct the regularization in Appendix D2 and verify the purification step.","tokens_in":24417,"tokens_out":48398,"duration_ms":419189,"concrete_test":"Implement the regularization of Appendix D2 for N=2 with qubit channels. Take a feasible Ω* from the noisy magnetometry SDP that produced Fig. 2(a), compute P* via condition (i), and construct Ω_ε and P_ε exactly as written in Eqs. (D23)-(D24) with ε = 0.1. Check (1) whether Tr_R[Ω_ε(|0⟩⟨0|_R ⊗ I_S)] = P_ε ⊗ I_OS, and (2) whether P_ε satisfies the sequential comb constraints of Eq. (A1). Under the published normalization, both checks fail. Then repeat with the identity term replaced by I_{S̄}/∏_{k=1}^{N} d_{2k−1}; if both checks pass and the optimal value of SP_k is unchanged as ε → 0, the concern is a typo. If no normalization yields a feasible P_ε, the converse inequality QP_k ≤ SP_k is unproven and the central claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The pivotal step is the converse direction of the central equality SP_k = QP_k in Appendix D2. To lift an arbitrary feasible Ω in SP_k to a feasible pair (P̃_ε, X_ε), the proof regularizes P* by adding an identity term. However, Eqs. (D23) and (D24) are mutually inconsistent: the added term in Ω_ε is ε/(∏_{k=1}^{2N} d_k) |0⟩⟨0|_R ⊗ I_S, while the claimed induced strategy operator is P_ε = (1−ε)P* + ε/(∏_{k=1}^{N} d_k) I_{S̄}. Taking the reference trace of condition (i) on Ω_ε yields (1−ε)P*⊗I_OS + ε/(∏_{k=1}^{2N} d_k) I_S, which matches P_ε⊗I_OS only if the two products coincide, which fails whenever the channel output dimensions exceed 1. Additionally, the identity I_{S̄} is not a feasible sequential comb under the stated normalization: for N=2, Tr_{H3}[I_{H1H2H3}] = d_3 I_{H1H2}, which is not of the form I_{H2}⊗P^{(1)} with Tr[P^{(1)}] = 1 unless the identity is normalized by d_1, not by the product. Thus Ω_ε does not belong to the feasible set of SP_k, and the constructed X_ε does not establish QP_k ≤ SP_k. The proof as printed therefore leaves the exact-SDP equivalence unproven. The gap is probably repairable by normalizing the identity with ∏_{k=1}^{N} d_{2k−1} (the input dimensions), which does make I_{S̄} a valid comb, but this correction is absent from the manuscript, and the numerical benchmarks do not test the regularization because the noiseless optimal combs are already full rank.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript develops a conic-programming framework for multiparameter quantum metrology, jointly optimizing probe preparation, intermediate controls, and measurements over parallel, sequential, and indefinite-causal-order strategy classes. The central technical claim is that the tight, Nagaoka-Hayashi, SLD, and Holevo precision bounds can be evaluated by an equivalent conic/SDP problem, with the exact equality SP_k = QP_k proved in Appendix D. The framework is benchmarked against known analytical results for noiseless multiparameter magnetometry (Eqs. (38)-(39) and Appendix F), extended to energy-constrained strategies, and specialized to finite-memory sequential schemes with identical controls. The paper concludes that a strict hierarchy among strategy classes can be identified in the noisy multiparameter regime, with causal superposition nearly matching sequential strategies and outperforming parallel and quantum-SWITCH schemes.","tokens_in":24833,"tokens_out":9705,"duration_ms":81764,"significance":"The contribution is potentially significant: if the exactness result holds, it provides a unified computational tool for determining achievable multiparameter precision limits across qualitatively different causal structures, going beyond the state-estimation conic framework of Ref. [50]. The paper deserves credit for benchmarking the SDP results against independent analytical bounds for parallel and sequential magnetometry, and for explicitly reproducing the optimal probe and control structures in Appendix F. The numerical evidence in Fig. 2(c) is consistent with the formulation, and the resource-constrained and finite-memory extensions address experimentally relevant scenarios. However, the exactness proof is not machine-checked and, as detailed below, contains a gap in the regularization step of the converse direction; the significance of the central claim depends on repairing that step.","major_comments":[{"comment":"Applying condition (i) in Eq. (21) to Ω_ε in Eq. (D23) gives Tr_R[Ω_ε(|0><0|_R⊗I_S)] = (1−ε)P*⊗I_OS + ε/(∏_{k=1}^{2N} d_k) I_S, whereas the claimed strategy operator in Eq. (D24) satisfies P_ε⊗I_OS = (1−ε)P*⊗I_OS + ε/(∏_{k=1}^{N} d_k) I_S. These agree only when ∏_{k=1}^{2N} d_k = ∏_{k=1}^{N} d_k, which is generically false. Moreover, I_{S_bar}/(∏_{k=1}^{N} d_k) is not a valid sequential comb under Eq. (A1): for N=2, Tr_{H3}[I_{H1H2H3}]/d1d2 = (d3/d1d2) I_{H1H2}, which is not of the required form I_{H2}⊗P^{(1)} with Tr[P^{(1)}]=1 unless d2=d3. The correct normalizing factor would be ∏_{k=1}^{N} d_{2k-1}. Consequently the proof of QP_k ≤ SP_k (Eq. (D38)) is incomplete as written; the numerical benchmarks in Fig. 2(c) do not exercise the regularization and cannot substitute for this argument.","section":"Appendix D2, Eqs. (D23)-(D24)"},{"comment":"The lift of Ω_ε to a purified strategy is under-specified. The text defines U as a unitary from H_S to H_OM, but the spectral decomposition in Eq. (D25) is of P_ε, which acts on S_bar, not on H_S; the congruence in Eq. (D28) needs U to map S_bar to H_OM with I_{R,OS} acting on the remaining system factor. Please clarify the domains of U and P_ε^{-1/2}; as written, the constructed X_ε is not well-defined.","section":"Appendix D2, Eqs. (D27)-(D29)"}],"minor_comments":[{"comment":"The implication symbols in these equations are corrupted ('/Leftr...'), making the logical direction of the cone-preservation claims difficult to verify; please repair the typesetting and re-check each implication.","section":"Eqs. (D11), (D30), (D31)"},{"comment":"The notation QP_k is reused for the finite-memory and identical-control problems, whereas QP_k already denotes the original conic problem in Section III; please disambiguate the notation.","section":"Section V"},{"comment":"The references to 'Fig. 1(a)' and 'Fig. 1(b)' in the discussion of Fig. 2 should be 'Fig. 2(a)' and 'Fig. 2(b)'.","section":"Main text after Fig. 2"},{"comment":"There are numerous typographical and spacing errors in the introduction (e.g., 'thesuccessinthesingle-parameterdomain', 'weconsider...'), which should be corrected before publication.","section":"Introduction"},{"comment":"The displayed matrix in Eq. (F9) appears to have formatting errors with missing or superfluous entries; please check the matrix against the analytical construction.","section":"Appendix F, Eq. (F9)"},{"comment":"The phrase 'the physical implementation of a general ICO is untraceable' is unclear; if 'untraceable' is intended to mean 'not realizable by a finite circuit', please state that explicitly.","section":"Section III"}],"recommendation":"major_revision","confidential_remarks":"The manuscript acknowledges overlapping concurrent work (Refs. [73,74]); the editor may wish to check how the present contribution is distinguished beyond the note added. The Appendix D gap is localized and likely repairable, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the bottom line: this paper deserves a serious referee, but not as-is. The central equivalence proof in Appendix D has a normalization error in the regularization step, so the converse direction of SP_k = QP_k does not go through as written. The error looks local and repairable, and the rest of the paper is genuinely useful.\n\nWhat is actually new: the authors extend Hayashi and Ouyang's conic programming from state estimation to channel estimation over quantum combs, giving exact SDP formulations for the Holevo, Nagaoka–Hayashi, and SLD bounds across parallel, sequential, and physically realizable indefinite-causal-order subclasses. The energy-budget and finite-memory extensions are new and practically motivated. The noiseless benchmarks are a real check: the SDP recovers the known analytical bounds of Eqs. (38)–(39) and the optimal probes and controls in Appendix F, which is solid evidence that the formulation is correct in at least the tested regime.\n\nThe soft spot is in Appendix D2. Equation (D23) adds ε/(∏_{k=1}^{2N} d_k) |0⟩⟨0|_R ⊗ I_S, while Eq. (D24) defines P_ε with denominator ∏_{k=1}^{N} d_k. Taking condition (i) on Ω_ε gives (1−ε)P*⊗I_OS plus ε/(∏_{k=1}^{2N} d_k) I_S, which matches P_ε⊗I_OS only if the two products coincide — generically false. Separately, the identity on ar S is not a valid sequential comb under the stated normalization; for N=2, Tr_{H3} I_{H1H2H3} = d_3 I_{H1H2}, which is not of the form I_{H2}⊗P^{(1)} with Tr[P^{(1)}]=1. Normalizing by the product of input dimensions, d_1 d_3 … d_{2N-1}, fixes both issues, so the proof is probably salvageable. But as printed, the exact-SDP equivalence is unproven. The numerical benchmarks do not test the regularization — the noiseless optimal combs are already full rank — so they do not rescue this step.\n\nOther issues are minor by comparison: no code is shipped, most figures lack solver tolerances and error bars, the abstract's \"strict hierarchy\" is stated more generally than the parameter-dependent examples support, and the finite-memory algorithm is acknowledged to be unstable at high noise. The citation pattern is fine, and the Note added honestly discloses overlapping concurrent work.\n\nWho this is for: anyone doing numerical multiparameter metrology or quantum-comb optimization. I would send it to review, with a request to fix the Appendix D normalization, make code and data available, and soften the hierarchy claim. The core framework is valuable enough to merit the referee time.","headline":"A useful extension of conic programming to multi-parameter channel metrology, with a localized but load-bearing proof bug in Appendix D that is likely repairable.","tokens_in":25405,"tokens_out":5544,"would_cite":true,"duration_ms":42169,"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":"The paper claims that, for parallel, sequential, and physically realizable indefinite-causal-order strategies, the optimal multiparameter estimation error is exactly computable as a semidefinite program whose feasible cones reproduce the…","keywords":["multiparameter quantum metrology","semidefinite programming","conic optimization","Holevo bound","Nagaoka-Hayashi bound","SLD bound","quantum combs","indefinite causal order"],"falsifier":"For a small channel (say two qubits, two parameters, and $N=2$ channel uses), solve the SDP and then grid-search over finite POVMs and locally unbiased estimators; if any explicit strategy beats the SDP value, the claimed equality $\\mathrm{SP}_k=\\mathrm{QP}_k$ fails. A second test is to check whether every feasible $\\Omega$ satisfying the cone and normalization conditions admits a purification through Eqs. (D21)-(D29); a feasible $\\Omega$ with no physical lift would also refute the proof.","tokens_in":24169,"feed_emoji":"⚛️","tokens_out":7055,"duration_ms":65526,"temperature":0.7,"pith_summary":"Estimating several parameters at once forces a tradeoff, because the best measurement for one parameter usually hurts another. This paper claims that the joint optimization of probe state, controls, and measurement can nonetheless be solved exactly for large classes of quantum strategies: the precision bounds become semidefinite programs whose optimal value provably equals the true minimum weighted estimation error. Different choices of the feasible cone give the tight, Nagaoka-Hayashi, Holevo, and SLD bounds, so one framework computes each bound and reveals how strategy class and resources affect precision. If the claim holds, experimenters can compute fundamental precision limits and optimal protocols for realistic multiparameter sensing directly, without heuristic searches.","feed_headline":"Precision limits for multiparameter quantum sensing computed exactly","feed_subtitle":"A single optimization over probes, controls, and measurements now yields the tightest known bounds.","key_machinery":"The central object is the auxiliary operator $X=\\sum_x v_x v_x^{\\mathsf T}\\otimes M_x$, built from the estimator deviations $v_x=(1,\\Delta\\hat\\theta(x))$ and the POVM elements $M_x$ on a tensor product of a real reference space $\\mathbb{R}^{d+1}$ with the strategy output space. It encodes the estimator and measurement in one variable, making the weighted mean-square error a linear trace objective. The strategy $\\widetilde P$ is then absorbed into a contracted variable $\\Omega=(I_R\\otimes\\Phi_{\\widetilde P})(X)$, where $\\Phi_{\\widetilde P}$ is the completely positive map induced by the quantum comb; this removes the bilinear coupling between strategy and measurement. Different precision bounds correspond to different cones $C_k$ on $X$: separable (tight), block-constrained (Nagaoka-Hayashi and SLD), and additionally antisymmetry-constrained (Holevo). The load-bearing step is the proof that these cones survive the strategy contraction and the reverse purification, yielding the equality $\\mathrm{SP}_k=\\mathrm{QP}_k$.","core_discovery":"For strategies with definite causal order, the paper proves in Appendix D that the conic relaxation is exact: the semidefinite-program optimum $\\mathrm{SP}_k$ equals the original optimization $\\mathrm{QP}_k$ over all probe states, control operations, and measurements, for the cones $k=2,3,4$ corresponding to the Nagaoka-Hayashi, SLD, and Holevo bounds, and also for the separable cone $k=1$ that gives the tight bound. The proof works by a strategy contraction that absorbs the quantum-comb strategy into the measurement-estimator operator, and then by a purification step that lifts every feasible contracted operator back to a physical strategy and measurement. Indefinite-causal-order strategies that decompose into definite-order sectors, such as causal superposition and the quantum switch, fit inside the same framework. The paper demonstrates the framework on three-parameter magnetometry, reproducing known analytical bounds in the noiseless case and identifying precision hierarchies under noise and energy constraints.","pith_inferences":["If the conic contraction is as general as the proof suggests, the same $\\Omega$ construction should transfer to channel discrimination and Bayesian quantum metrology, problems the paper itself names as future directions.","The reported ordering (causal superposition and sequential strategies nearly tied, both outperforming parallel and quantum switch) is demonstrated for a particular magnetometry model; treating it as a universal hierarchy would be an extrapolation beyond the paper.","Because the tight bound's separable cone is not semidefinite representable, the exact SDPs for the other three cones can serve as certified bounds from which to sandwich the tight bound in finite-sample settings.","The finite-memory identical-control iteration is nonconvex and occasionally fails to converge at high noise; a natural test is whether updating several control positions per step, rather than one randomly chosen site, improves convergence without increasing memory."],"forward_implications":["The same SDP computes the exact Holevo, Nagaoka-Hayashi, and SLD bounds for parallel and sequential strategies with $N$ channel uses, jointly optimizing probe state, controls, and measurement.","Physically realizable indefinite-causal-order protocols, including causal superposition and the quantum switch, can be included in the optimization, so their multiparameter precision can be compared with definite-order strategies within one framework.","Energy-budget constraints and finite-memory restrictions become linear or decomposable constraints on the optimization variable, allowing resource-limited sensing protocols to be optimized rather than searched heuristically.","In the noiseless magnetometry model, the numerical bounds reproduce the known analytical parallel and sequential limits, supporting the claim that the SDP optimum is the true precision bound.","Resource-constrained comparisons quantify the precision loss from limited energy and memory, giving a practical route from ideal limits to near-term sensor design."],"supporting_citations":[{"why":"Supplies the conic-programming approach to multiparameter probe optimization that this work extends from state estimation to channel estimation.","marker":"[48]"},{"why":"Provides the definitions of the separable, Nagaoka-Hayashi, SLD, and Holevo cones and the hierarchy among the corresponding bounds.","marker":"[50]"},{"why":"Defines the Nagaoka-Hayashi bound for separable measurements, one of the central bounds reproduced by the framework.","marker":"[45]"},{"why":"Gives the analytical sequential feedback bound and optimal control used to validate the numerical method in the noiseless regime.","marker":"[15]"},{"why":"Gives the analytical parallel-strategy precision limit and optimal probe state used as the noiseless parallel benchmark.","marker":"[35]"},{"why":"Provides the minimal-ancilla realization theorem that converts an optimal comb into an explicit quantum circuit.","marker":"[59]"},{"why":"Defines the global-battery energy-constraint model on which the resource-constrained formulation is built.","marker":"[69]"},{"why":"Introduces the iterative multi-convex optimization method adapted for finite-memory sequential strategies.","marker":"[70]"}],"fun_headline_variants":["Exact multiparameter sensing bounds via SDP","One optimization: tightest precision bounds for multiparameter metrology","SDP solves multiparameter quantum sensing limits","Framework yields exact multiparameter metrology bounds","Exact limits for multiparameter quantum estimation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole construction rests on the Appendix D step that every feasible compressed strategy can be rebuilt as a physical strategy-and-measurement pair; the paper gives a proof sketch for this step but no independent or machine-checked verification.","fun_headline_variants_meta":{"raw":{"variants":["Exact multiparameter sensing bounds via SDP","One optimization: tightest precision bounds for multiparameter metrology","SDP solves multiparameter quantum sensing limits","Framework yields exact multiparameter metrology bounds","Exact limits for multiparameter quantum estimation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000817,"raw_usage":{"total_tokens":3592,"prompt_tokens":970,"completion_tokens":2622,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":586,"completion_tokens_details":{"reasoning_tokens":2546}},"tokens_in":586,"tokens_out":2622,"duration_ms":15624,"temperature":1.0,"reasoning_tokens":2546,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:12:37.750892+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a small channel (say two qubits, two parameters, and $N=2$ channel uses), solve the SDP and then grid-search over finite POVMs and locally unbiased estimators; if any explicit strategy beats the SDP value, the claimed equality $\\mathrm{SP}_k=\\mathrm{QP}_k$ fails. A second test is to check whether every feasible $\\Omega$ satisfying the cone and normalization conditions admits a purification through Eqs. (D21)-(D29); a feasible $\\Omega$ with no physical lift would also refute the proof.","supporting_citations":[{"cited_title":"Hayashi and Y","cited_arxiv_id":null,"evidence_quote":"Provides the definitions of the separable, Nagaoka-Hayashi, SLD, and Holevo cones and the hierarchy among the corresponding bounds."},{"cited_title":"Yuan, Sequential feedback scheme outperforms the parallel scheme for hamiltonian parameter estimation, Physical review letters117, 160801 (2016)","cited_arxiv_id":null,"evidence_quote":"Gives the analytical sequential feedback bound and optimal control used to validate the numerical method in the noiseless regime."},{"cited_title":"Bisio, G","cited_arxiv_id":null,"evidence_quote":"Provides the minimal-ancilla realization theorem that converts an optimal comb into an explicit quantum circuit."},{"cited_title":"Chen and Y","cited_arxiv_id":null,"evidence_quote":"Defines the global-battery energy-constraint model on which the resource-constrained formulation is built."},{"cited_title":"Liu and Y","cited_arxiv_id":null,"evidence_quote":"Introduces the iterative multi-convex optimization method adapted for finite-memory sequential strategies."}],"review_version":1}