{"id":"b0c929ed-6aeb-45d6-b87b-cbbfdfa43aa9","arxiv_id":"2601.23043","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Dicke superposition probes achieve near-Heisenberg (linear) and super-Heisenberg (two-body) Fisher-information scaling and show improved phase-damping robustness relative to GHZ and W-superposition states under a pre-encoding noise model.","lead":"The paper studies multi-qubit quantum states called Dicke superpositions as sensors of tiny phase shifts, and claims they resist certain kinds of noise better than standard entangled states. A generalist might care because these states are among the easier entangled states to build, so they are a practical candidate for noisy quantum sensors.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Noise is applied before encoding, so the central 'noise-resilient phase sensing' claim is not established for decoherence during interrogation.","rationale":"The reader's weakest-assumption analysis correctly identifies the pre-encoding noise model as the load-bearing weakness. This is not a flaw in the noiseless QFI derivations or in the numerical evaluation of F_Q(Λ(ρ),H); it is a mismatch between the model analyzed and the central claim of noise-resilient phase sensing. The proposed concrete test would settle whether the robustness ordering survives during-encoding decoherence. Since the paper can be fixed by scope clarification or by adding the standard during-encoding calculation, the conditional verdict remains appropriate rather than a rejection. Secondary internal inconsistencies (e.g., Table II's λ_min for H_2^{(3)} with odd N appears inconsistent with the paper's own FQ formula, and Table I's l ranges are not fully consistent with Eq. (22)) reinforce the need for revision but are not the primary threat to the central claim.","tokens_in":17856,"tokens_out":6647,"duration_ms":69940,"concrete_test":"Recompute the N=8 phase-damping curves (Fig. 3a and Fig. 6) with decoherence during encoding: discretize U(θ)=e^{-iHθ} into M segments, apply the local phase-damping channel Λ_p^{⊗N} after each segment, and compute the QFI of the final state at θ for |D_{3,5}^{(8)}⟩, GHZ, and |D_{4,4}⟩. Equivalently, solve dρ/dt = -i[H,ρ] + γ Σ_i (σ_z^{(i)} ρ σ_z^{(i)} - ρ) and evaluate F_Q at the same θ. If the enhanced robustness of the Dicke superposition relative to GHZ (and the similarity to balanced Dicke) disappears or reverses, the central claim must be restricted to preparation noise.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central noise-resilience claim rests on a specific noise ordering: Sections III.B and IV.B evaluate ρ_noisy via Eq. (25) or (28), and then compute the QFI using Eq. (8). That is equivalent to F_Q(Λ(ρ), H) for a noiseless unitary encoding U(θ)=e^{-iHθ}, i.e., state-preparation noise. In standard noisy quantum metrology, the relevant scenario is decoherence during the interrogation, where the channel acts while the parameter is being encoded; the QFI is not generally F_Q(Λ(ρ), H), and the relative ordering of probes can change. For phase damping in particular, GHZ-type states are known to lose their advantage under during-encoding dephasing, so the apparent robustness of |D_{3,5}^{(8)}⟩ in Figs. 2–3 and Fig. 6 is computed in a model that does not match the usual sensing setup. Therefore the abstract's claim that these states are 'noise-resilient resources' under 'realistic decoherence channels' is not established. The noiseless QFI calculations are unaffected, and the paper could be repaired either by explicitly scoping the claims to preparation noise or by adding a during-encoding analysis.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the quantum Fisher information (QFI) and phase sensitivity of N-qubit Dicke-state superposition probes under unitary encodings generated by one-body Hamiltonians H = J_n and by two-body Hamiltonians J_n^2, J_n + J_n^2, and their regularized variants (Eq. 29). In the noiseless case it shows that certain Dicke superpositions reach QFI ~ (3/4)N^2 for linear encoding (Eqs. 23–24) and that the optimal pure probes for the two-body Hamiltonians are equal superpositions of the extremal eigenstates, giving F_Q = (λ_max − λ_min)^2 (Table II). The paper then applies local phase-damping, amplitude-damping, and global-depolarizing channels to the probe state before encoding, computes the QFI of the resulting noisy state under a unitary evolution, and compares the robustness of the proposed probes with GHZ, W-superposition, and balanced Dicke states. The central claim is that tailored near-optimal Dicke superpositions are versatile, noise-resilient resources for Heisenberg and super-Heisenberg phase sensing.","tokens_in":18112,"tokens_out":11150,"duration_ms":97371,"significance":"If the results are taken as stated, the paper would identify experimentally relevant, permutation-symmetric probe families that combine near-optimal noiseless QFI with improved robustness under phase damping. The noiseless QFI derivations are largely self-contained and standard: Eq. (10) correctly bounds pure-state QFI by the squared spectral range, and the optimal states in Table II follow from the extremal eigenstates. The paper also usefully unifies linear and two-body metrology benchmarks (NL-SNL/NL-HL) and gives explicit analytic formulas for the two-body optimal probes. However, the noise-resilience claim is not established for the usual metrological scenario in which decoherence acts during the parameter encoding; the present analysis only treats state-preparation noise. There is also a concrete error in Table II for H_2^(3) with odd N, and the notation for the near-optimal Dicke pairs in Table I is inconsistent with the defining equation. These issues affect the reproducibility and the central advertised conclusion, but they are fixable within the scope of the manuscript.","major_comments":[{"comment":"The noise is applied to the probe state before the unitary encoding: ρ_noisy is computed from Eq. (25) or (28), and the QFI is then evaluated using Eq. (8) for the noiseless unitary U(θ)=e^{-iHθ}. This is a state-preparation-noise model, not decoherence during the interrogation. In standard noisy metrology the channel acts while the parameter is being encoded, and the QFI is not generally F_Q(Λ(ρ),H); the relative ordering of probes can change. For phase damping in particular, GHZ-type advantage is known to degrade under during-encoding dephasing, so the robustness advantage reported in Figs. 2–3 and 5–7 is computed in a model that does not match the usual sensing setup. The abstract and conclusions claim 'noise-resilient resources' under 'realistic decoherence channels'; as written, that claim is not established. Please either explicitly scope the claim to preparation noise or add a dur","section":"Sec. III.B, Eqs. (25)–(28); Sec. IV.B"},{"comment":"For odd N the table lists λ_min = −N/8 for H_2^(3) = (1/2)(J_n^2 − (N/4)I), but the associated eigenstate |D_{(N−1)/2,(N+1)/2}> has J_n eigenvalue −1/2, giving H_2^(3) eigenvalue (1/2)(1/4 − N/4) = (1−N)/8. The quoted F_Q = (N^2−1)^2/64 follows from λ_min = (1−N)/8, not from −N/8. The table entry should be corrected.","section":"Table II, row H_2^(3)"},{"comment":"The definitions are inconsistent. Equation (22) restricts l,l' to {1,...,N−2}, but the odd-N condition in Eq. (23), N/2 − l' = ±3/2, gives l' = (N∓3)/2, which for N=5 is l'=1 or 4, with 4 outside the allowed range. The entries in Table I such as (l,l')=(±2,0) and (±4,∓2) cannot be parsed as l values satisfying Eq. (22), and the caption claims these pairs 'maximize the QFI' without a maximization proof. Please clarify the notation (e.g., whether the table lists (M,M') or (l,l')), state the allowed ranges explicitly, and verify the entries against the formulas. This is needed to make the near-optimal probe family reproducible.","section":"Sec. III.A, Eqs. (22)–(24) and Table I"}],"minor_comments":[{"comment":"The maximization in the definition of QFI should be over POVMs; as written, 'max over ρ,E' is a typo.","section":"Eq. (5)"},{"comment":"The caption labels the optimal probe for H_2^(1) as |ψ^(2)_{(N=8)}> = (|GHZ>_n + |D_{4,4}>_n)/√2, but Eq. (30) shows that this is |ψ^(1)>, not |ψ^(2)>. The label should be corrected.","section":"Fig. 5 caption"},{"comment":"The caption says 'see Table III' for the near-optimal Dicke superposition states; the relevant table is Table IV.","section":"Fig. 6 caption"},{"comment":"The table lists numerical QFI values for near-optimal Dicke superpositions for N=5–8, but the text does not give the formulas or the optimization procedure used to obtain these values. A brief description or analytic expressions would improve reproducibility.","section":"Table IV"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a solid noiseless QFI analysis with an attractive Dicke-superposition construction, but the advertised noise-resilience result is currently computed for a preparation-noise model only, while the conclusions are phrased in terms of realistic decoherence. The Table II inconsistency and the Table I/Eq. (22) notation problem are concrete and should be fixed before publication. I recommend major revision rather than rejection because the noiseless calculations are sound and the noise claim can be repaired by re-scoping or by adding a during-encoding analysis."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The noiseless part of this paper is genuinely useful. The QFI formulas for Dicke superpositions under J_n and the four J^2-type two-body Hamiltonians are standard but correctly applied, and the near-optimal Dicke superpositions for the two-body encodings (Table IV) are new. The construction via extremal eigenstates is clean, and the scaling analysis is sound. I would happily cite the noiseless formulas and the two-body probe table.\n\nThe main thing you should know: the noise analysis applies the channel to the probe before the unitary encoding, then computes QFI for the noiseless evolution. That is preparation noise, not decoherence during interrogation. The abstract and conclusions call these states \"noise-resilient resources\" under \"realistic decoherence channels,\" which overstates what is shown. For phase damping, the relative ordering of probes can completely change when the channel acts during encoding, and that scenario is never analyzed. The robustness of |D^(8)_3,5> in Figs. 2-3 and Fig. 6 is computed in exactly this preparation-noise model. This is fixable by scoping the claims, but as written the central noise-resilience claim is not established for the standard noisy-sensing setup.\n\nThere are a few internal inconsistencies. Table II lists lambda_min(odd N) = -N/8 for H^(3)_2, but the actual minimum of (1/2)(J^2 - N/4 I) is (1-N)/8; the QFI entry corresponds to the correct value, so the table entry is wrong, not the final result. Table I uses pairs like (±2,0) and (±3,∓1) that fall outside the stated range l = 1, …, N-2 in Eq. (22); the range needs relaxing. Fig. 5's caption labels the probe as psi^(2) but the superposition shown is psi^(1). These are minor but should be fixed.\n\nThe paper's noiseless contributions are not affected by the noise-ordering issue. The QFI tables and the near-optimal state lists for two-body Hamiltonians are a convenient reference. The noise analysis is a cautionary example of how the choice of noise model matters, and the authors should be asked to either explicitly scope their claims to preparation noise or add a during-encoding analysis.\n\nThis deserves a serious referee. The core noiseless content is correct, the inconsistencies are easy repairs, and the noise-ordering point is an important methodological lesson. I would send it to peer review and ask for revision rather than desk reject.","headline":"Solid noiseless Dicke-superposition results, but the headline noise-resilience claims rest on a noise-before-encoding model that the paper never flags; the internal inconsistencies are minor but the noise-ordering issue is not.","tokens_in":794,"tokens_out":812,"would_cite":true,"duration_ms":43994,"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":"Superpositions of two Dicke states, already feasible in photonic chips, give near-optimal quantum Fisher information under linear and two-body collective-spin encoding and stay measurably more sensitive to dephasing than GHZ or W-superposit","keywords":["Dicke states","quantum metrology","quantum Fisher information","Heisenberg limit","super-Heisenberg scaling","phase damping","two-body interactions","collective spin"],"falsifier":"Compute the QFI under the same phase-damping, amplitude-damping, and depolarizing channels but with the channel applied during the unitary encoding (interleaved with small increments of e^{-iH dθ}) rather than before it. If the Dicke-superposition advantage over GHZ vanishes in this during-encoding model, the paper's noise-resilience claim is restricted to state-preparation noise.","tokens_in":17704,"feed_emoji":"⚛️","tokens_out":5255,"duration_ms":50054,"temperature":0.7,"pith_summary":"The paper claims that equal-weight superpositions of two Dicke states form a versatile, noise-resilient probe family for phase estimation under one-body and two-body collective-spin generators. For linear encoding, these near-optimal Dicke superpositions reach a quantum Fisher information (QFI) of roughly 3/4 N^2 and, for N=8, retain more phase sensitivity under phase damping than GHZ, W-superposition, and balanced Dicke states. For two-body interactions such as J_n^2 and J_n+J_n^2, the optimal pure probes are superpositions of the Hamiltonian's extremal eigenstates, achieving the spectral-bound QFI (lambda_max - lambda_min)^2, while near-optimal Dicke superpositions again show improved resilience to phase damping. These results matter because Dicke superpositions are experimentally accessible in on-chip photonic setups, making the predicted robustness a testable resource for practical quantum sensing.","feed_headline":"Dicke superpositions keep Heisenberg sensitivity under dephasing","feed_subtitle":"Superpositions of just two Dicke states preserve most of the N^2 sensitivity that GHZ states lose to phase noise.","key_machinery":"The central objects are equal-weight superpositions of two Dicke states, |D_{l,l'}^{(N)}> = (|D_{N−l,l}> + |D_{N−l',l'}>)/√2, where |D_{N−l,l}> is the symmetric N-qubit state with l excitations. The workhorse identity is the pure-state QFI formula FQ = 4(ΔH)^2, which for a superposition of extremal eigenstates reduces to (λmax−λmin)^2. For linear encoding H = J_n, choosing l,l' near N/2 yields FQ≈(3/4)N^2; for two-body Hamiltonians such as J_n^2 and J_n+J_n^2, the optimal probe is the equal superposition of the maximum and minimum eigenstates, and the paper evaluates the resulting super-Heisenberg scaling and compares the dephasing resilience of these optimal probes with near-optimal Dicke s","core_discovery":"The paper establishes that a specific family of permutation-symmetric states—equal superpositions of two Dicke states—provides near-optimal quantum Fisher information under both linear collective-spin encoding and two-body interaction encoding, and that these states retain more phase sensitivity under local phase damping than the standard GHZ and W-superposition probes. For linear generators, an analytic expression yields QFI ≈ (3/4)N^2 for the best choices of Dicke pairs. For two-body Hamiltonians, the paper derives the optimal probe as the equal superposition of the maximum and minimum eigenstates, with QFI equal to the squared spectral range (λmax−λmin)^2, and shows that near-optimal Dick","pith_inferences":["The robustness ordering found for state-preparation noise may not survive when decoherence acts during the interrogation; a fair experimental test should specify which noise timing is being probed.","A natural design principle suggested by the results: spreading the probe over the Dicke manifold averages out transverse dephasing in J_z, which could guide construction of larger-N probes beyond the explicit table entries.","The same extremal-eigenstate recipe may generalize to k-body generators with k>2, where superpositions of symmetric minimal and maximal eigenstates would give the spectral-range QFI while near-optimal variants could mitigate noise.","Because Dicke superpositions have already been generated on-chip for four photons, the predicted dephasing resilience could be tested directly on existing platforms by comparing |D_{3,5}> with GHZ and W states under controlled phase damping."],"forward_implications":["Near-optimal Dicke superpositions reach FQ ≈ 0.75 N^2 for linear collective-spin encoding, close to the Heisenberg limit N^2, for both odd and even N.","For two-body generators, probes that superpose the extremal eigenstates attain the spectral bound FQ = (λmax−λmin)^2, which scales as N^4 for J_n^2—super-Heisenberg phase sensitivity.","For N=8, the |D_{3,5}> state retains higher QFI than GHZ, W-superposition, and balanced Dicke states under phase damping, and comparable behavior under global depolarization.","Near-optimal Dicke superpositions outperform the QFI-optimal two-body probes under local phase damping for the system sizes studied, giving a practical trade-off between ideal sensitivity and noise resilience.","The benchmarks in the paper's Table III provide concrete reference values (NL-SNL and NL-HL) against which any N=8 probe under J_n^2-type encodings can be judged."],"fun_headline_variants":["Dicke superpositions beat GHZ under dephasing","Two Dicke states keep Heisenberg scaling in noise","Dicke pairs dodge dephasing for quantum sensing","Superposed Dicke states outlast GHZ in phase noise"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The noise analysis places the decoherence channel before the unitary phase encoding and then computes the QFI under noiseless evolution, so the predicted robustness is for preparation noise, not for decoherence that occurs while the phase is being imprinted.","fun_headline_variants_meta":{"raw":{"variants":["Dicke superpositions beat GHZ under dephasing","Two Dicke states keep Heisenberg scaling in noise","Dicke pairs dodge dephasing for quantum sensing","Superposed Dicke states outlast GHZ in phase noise"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000148,"raw_usage":{"total_tokens":1015,"prompt_tokens":725,"completion_tokens":290,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":469,"completion_tokens_details":{"reasoning_tokens":223}},"tokens_in":469,"tokens_out":290,"duration_ms":3707,"temperature":1.0,"reasoning_tokens":223,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T06:15:02.843841+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the QFI under the same phase-damping, amplitude-damping, and depolarizing channels but with the channel applied during the unitary encoding (interleaved with small increments of e^{-iH dθ}) rather than before it. If the Dicke-superposition advantage over GHZ vanishes in this during-encoding model, the paper's noise-resilience claim is restricted to state-preparation noise.","supporting_citations":[],"review_version":1}