{"id":"279eb7de-8c0a-4d38-b0b6-f9c59505d14d","arxiv_id":"2512.24406","paper_version":4,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"In Heisenberg-coupled qubit arrays, Dicke states including W states can be prepared with a single local control in times that grow approximately quadratically with qubit number (numerically up to N=9).","lead":"This paper shows numerically that Dicke states — special entangled states of many qubits — can be generated in a chain of up to nine qubits by steering just one qubit with a single time-varying magnetic field. The result matters because it points to a minimal-control route to preparing useful entangled states in quantum devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'shortest-possible' scaling claim is not certified: dCRAB multistart finds upper bounds, not proven minima, and the N=3..9 fits have no uncertainty quantification.","rationale":"The paper is a solid numerical application of known subspace controllability, but its headline physical claim is 'shortest possible times scale quadratically.' That claim requires each reported T_min to be a true global minimum of the infidelity landscape. The manuscript provides no certificate and describes only a heuristic global search; the stability check mentioned in Sec. V C is not a proof. The reader's weakest-assumption identification is exactly this. I also note the abstract/text exponent discrepancy and lack of error bars on the fits, which compound the uncertainty but do not by themselves invalidate the approach. These are addressable issues rather than fatal flaws, so the CONDITIONAL verdict remains appropriate without modification.","tokens_in":21542,"tokens_out":3757,"duration_ms":45171,"concrete_test":"Reproduce the time-optimal search for N=3,4,5,6 with an independent method: use piecewise-constant controls optimized by GRAPE or Krotov with thousands of random seeds, and for N=4 (a=1,2 subspace dimensions 4 and 6) run a certified global search (e.g., branch-and-bound or dense grid + local refinement) over the Fourier coefficients for the same T grid. If any control with 1-F<10^-3 is found at T below the reported T_min, the 'shortest possible' claim collapses. If all methods reproduce the same T_min and no shorter T exists, the claim gains support. Also report bootstrap confidence intervals for the power-law exponents and test N=10 if feasible.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim — that the shortest possible times scale quadratically with N (Sec. VI, Fig. 6) — depends on the premise that the dCRAB multistart search finds the global infidelity minimum at each T. Section V C describes sampling ~10^3 points, selecting ~20, running Nelder-Mead, and 'adopting' the smallest as 'the desired global minimum,' with only sample-size stability as validation. There is no global-optimality certificate, no comparison with an independent optimizer, and no error bar on any T_min. Every T_min is therefore an upper bound on the true shortest time; a better pulse at shorter T would lower the fitted exponents or change their form. The abstract/full-text exponent mismatch (O(N^2.08)/O(N^1.78) vs 0.12 N^2.02/0.14 N^2.17) also signals unreported variability in the fits. With only N=3..9, the quadratic extrapolation to larger N is not quantitatively supported. These are correctness risks, not internal inconsistencies in the controllability argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies Dicke-state generation in a linear qubit array with always-on isotropic Heisenberg nearest-neighbor coupling and a single local Zeeman-type Z control on one actuator qubit. It invokes an existing Lie-algebraic subspace-controllability result to argue that any Hamming-weight-a product state can be steered to the a-excitation Dicke state, and then uses a dCRAB-based optimal-control search (with multistart clustering and Nelder-Mead local optimization) to find smooth control pulses that achieve target-state infidelity below 10^-3 for arrays up to N=9. The reported shortest preparation times are fitted as T_min ≈ 0.12 N^{2.02} for W states and T_min ≈ 0.14 N^{2.17} for a=2 Dicke states, leading the authors to claim that the shortest possible state-preparation times scale quadratically with N. The paper also reports an analysis of robustness to errors in the pulse-expansion coefficients.","tokens_in":21908,"tokens_out":4325,"duration_ms":47641,"significance":"If the scaling claim were certified, the result would be of clear interest: it would show that a minimal control resource—a single local Z field on a Heisenberg chain—can generate highly entangled Dicke states in times growing only quadratically with system size, and it would complement existing Lie-algebraic controllability results with concrete pulse-level constructions. The numerical study is carefully described: the algorithm parameters, fidelity threshold, and state-evolution method are specified, and the reported infidelities are direct Schrödinger-evolution results. The robustness analysis, though limited, addresses a practical concern. The main weakness is that the word 'shortest possible' is not supported by the numerical evidence: the dCRAB/multistart procedure is a heuristic global-search method without a global-optimality certificate, so the reported T_min values are upper bounds. In addition, the abstract's scaling exponents do not match the fitted exponents in Sec. VI, and the fit uses only seven points with no uncertainty quantification.","major_comments":[{"comment":"The central claim—'shortest possible state-preparation times'—is load-bearing in the abstract and in Sec. VI, but it is not supported by the optimization methodology. The multistart/dCRAB procedure described in Sec. V C samples ~10^3 random points, keeps ~20, runs Nelder-Mead local searches, and adopts the lowest minimum as 'the desired global minimum' based on stability under changing the sample size. This is a heuristic without a global-optimality certificate, and no independent optimizer (e.g., GRAPE/Krotov) is used for cross-validation. Every reported T_min is therefore an upper bound on the true minimal time; a better control pulse at shorter T would lower the fitted exponents. The dependence on the chosen amplitude bound B_max = 4π J is also not explored. Please either (i) soften the claim to 'shortest times found by this search' and revise the abstract and Sec. VI accordingly, or","section":"Sec. V C and Sec. VI"},{"comment":"The abstract states that the shortest times scale as O(N^{2.08}) for W states and O(N^{1.78}) for a=2 Dicke states, while Sec. VI and Fig. 6 report fitted curves T_min = 0.12 N^{2.02} and T_min = 0.14 N^{2.17}. These are not minor rounding differences: the exponents differ by 0.06 and 0.39, respectively. Since the quantitative scaling is the paper's principal conclusion, this inconsistency must be resolved. Furthermore, the fit uses only N=3,...,9 (seven points), with no confidence intervals, residuals, or sensitivity analysis; the extrapolation to 'scale quadratically with N' is therefore not quantitatively established beyond the fitted range.","section":"Abstract vs. Sec. VI; Fig. 6"},{"comment":"The robustness analysis perturbs the expansion coefficients c_m^(l) and s_m^(l), not the actual control field B(t). The abstract and Sec. VII claim robustness against 'small control-field deviations from the optimal values,' which is a stronger statement. Because B(t) is a sum of many oscillatory Fourier terms, a 5% error in an individual coefficient does not imply a 5% bound on the pointwise field B(t) or on its time derivative. Please either test direct additive/multiplicative field noise, or reformulate the robustness conclusion as being specifically about the pulse parametrization coefficients.","section":"Sec. VI, robustness subsection (Figs. 11-12)"}],"minor_comments":[{"comment":"Typo: 'neigbor' should be 'neighbor' in the opening sentence.","section":"Sec. VII"},{"comment":"The penalty term in the figure of merit is discontinuous (linear in the violation via a Heaviside factor). Since the local optimizer is Nelder-Mead, a nonsmooth objective may affect convergence; a brief remark or a smooth quadratic penalty would be preferable.","section":"Eq. (26)"},{"comment":"The statement that the results depend on M 'only in an implicit fashion' is made without numerical evidence. A short M-scan (e.g., M=10,15,20 for one state) would make this claim verifiable.","section":"Sec. V C"},{"comment":"The caption refers to 'the highest fidelities achieved,' while the text says infidelities are mostly between 10^-3 and 10^-4. Please clarify whether the plotted quantity is fidelity or infidelity.","section":"Sec. VI, Fig. 5 caption"}],"recommendation":"major_revision","confidential_remarks":"To the editor: the manuscript would benefit from releasing the raw T_min values and pulse parameters, since the central scaling claim is computational and currently not independently reproducible. The abstract/full-text exponent mismatch should be resolved before acceptance; the 'shortest possible' wording should also be brought in line with what a heuristic optimizer can actually certify."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a competent numerical study of Dicke-state generation in Heisenberg chains with a single local Z control. The new content is the dCRAB optimal-control results: explicit pulses, T_min values, robustness analysis, and fitted scalings. The subspace-controllability theorem is prior work (ref [40]), and the paper is upfront about that. Credit where due: the model is clean, the dCRAB implementation with multistart clustering is careful, and the robustness analysis (5% parameter errors give ~1e-4 fidelity loss) is genuinely useful for experimentalists. The fitted quadratic scaling from N=3..9 is plausible and a good benchmark for the field.\n\nThe soft spots are real but addressable. First, the phrase 'shortest possible time' is not supported. dCRAB is a heuristic; the multistart procedure samples ~10^3 points, keeps ~20, and adopts the smallest local minimum without any global-optimality certificate. Every T_min is an upper bound, not a proven minimum. The paper should say 'shortest time found by our method' or 'near-optimal.' Second, the abstract reports exponents O(N^2.08) and O(N^1.78), while the full text gives 0.12 N^2.02 and 0.14 N^2.17. That is a serious internal inconsistency and signals the fits are fragile. Third, with only seven points and no uncertainty quantification, the extrapolation to larger N is not quantitatively supported. Fourth, lack of code/data prevents independent verification. Minor point: the robustness test only perturbs Fourier coefficients, not physical parameters like J or the pulse amplitude calibration.\n\nNone of this is fatal. The controllability argument is solid, the numerics are reproducible in principle, and the quadratic scaling is a useful empirical result even if not a proven lower bound. This paper is for quantum-control practitioners and solid-state experimentalists looking for minimal-resource routes to Dicke and W states. It deserves serious peer review, provided the authors fix the exponent inconsistency, soften the 'shortest possible' language, and ideally share code. I would send it to a referee rather than desk-reject.","headline":"The genuinely new part is the dCRAB numerics, not the controllability theorem, but the 'shortest possible' claim and the abstract's scaling exponents both overreach what the optimization actually certifies.","tokens_in":22348,"tokens_out":1824,"would_cite":true,"duration_ms":20540,"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 single local Z control suffices to generate Dicke states in Heisenberg chains with minimal times scaling quadratically with qubit number.","keywords":["Dicke states","W states","subspace controllability","Heisenberg coupling","single local control","quantum optimal control","dCRAB","time-optimal state preparation"],"falsifier":"For any N in 3–9, run a more exhaustive search (e.g., M ≥ 30 harmonics, multiple global optimizers, or a grid of T below the reported T_min) and check whether a control field with the same fidelity threshold 1−10^-3 exists. If one is found, the claim of shortest-possible time fails. Alternatively, an analytic lower bound on T_min growing faster than N^2 would falsify the extrapolated scaling.","tokens_in":21456,"feed_emoji":"⚛️","tokens_out":6711,"duration_ms":60658,"temperature":0.7,"pith_summary":"The paper establishes that Dicke states — symmetric equal-weight superpositions of all bit strings with a fixed number of excitations — can be prepared in a Heisenberg-coupled qubit chain using just one local Z control on a single actuator qubit, starting from any product state with the same excitation number. The theoretical basis is subspace controllability: the excitation-number symmetry decomposes the Hilbert space into invariant fixed-Hamming-weight subspaces, and on each such subspace a single local control is provably sufficient to reach any state. Using the dCRAB optimal-control algorithm with a multistart global search, the authors find numerically that the shortest preparation times for W states and two-excitation Dicke states scale close to quadratically with N for arrays up to 9 qubits, with fitted curves T_min = 0.12 N^{2.02} (W) and T_min = 0.14 N^{2.17} (a=2). If correct, this gives a near-minimal-control route to a practically important family of entangled states, with favorable scaling compared to other analog schemes.","feed_headline":"Quadratic time to Dicke states with one local control","feed_subtitle":"Subspace controllability lets a Heisenberg chain make entangled states without per-qubit addressing.","key_machinery":"The key objects are (i) the excitation-number symmetry S_exc = (1/2) Σ(1+Z_n), which commutes with both the Heisenberg drift Hamiltonian and the local Z control, decomposing the Hilbert space into invariant subspaces of fixed Hamming weight; and (ii) the dressed Chopped Random Basis (dCRAB) algorithm, which optimizes smooth control fields in a truncated random Fourier basis with dressing iterations and multistart clustering. Together they turn a Lie-algebraic existence guarantee into concrete, near-time-optimal pulses: the symmetry ensures any same-weight state is reachable, while dCRAB locates the shortest duration T at which the target fidelity exceeds 1−10^-3.","core_discovery":"The central claim is that the minimal time needed to evolve a fixed-Hamming-weight product state into the corresponding Dicke state under the isotropic Heisenberg Hamiltonian plus a single local Z field grows quadratically with the number of qubits N. The paper proves reachability from Lie-algebraic subspace controllability on each excitation subspace, then supplies numerical optimal-control evidence: for W states (a=1) and a=2 Dicke states in arrays of 3–9 qubits, the fitted minimal times are T_min = 0.12 N^{2.02} and T_min = 0.14 N^{2.17}, with infidelities between 10^-3 and 10^-4 at these times; the abstract quotes exponents 2.08 and 1.78, both consistent with a quadratic law. The authors","pith_inferences":["An editorial extension left implicit by the paper: the fitted exponents for N=3..9 are treated as indicators of an asymptotic quadratic law, but the true asymptotic exponent could be slightly different; a testable prediction is that the a=2 exponent remains at or below the W-state exponent at larger N.","A natural extension is to other coupling geometries and actuator placements; one could test whether the quadratic scaling persists when the actuator sits at an interior qubit or when the chain becomes a ring.","The minimal-time scaling may reflect a Lie-algebraic speed limit set by the control Hamiltonian's norm on each subspace; deriving an analytic lower bound of order N^2 would convert the numerical finding into a theorem."],"forward_implications":["If the quadratic scaling holds for larger N, Dicke and W states can be prepared in times of order N^2/J without individual qubit addressing, improving on superlinear analog-scheme scalings.","The single-control, always-on-interaction setting is compatible with spin-based qubit arrays where global addressing is hard; the smooth bounded pulses are within reach of arbitrary waveform generators.","The subspace-controllability argument applies to any final state with the same Hamming weight as the initial state, so the same machinery can target other fixed-weight entangled states, not only Dicke states.","The reported robustness to 5% parameter errors suggests that the scheme tolerates realistic control-field imperfections in experiment."],"fun_headline_variants":["Single control, quadratic time to entangled states","One knob steers Heisenberg chain to Dicke states in ~N^2 time","Dicke states with a single local pulse: time grows as N^2","Subspace control: reach Dicke states in quadratic time with one qubit","From product to Dicke: one local Z, quadratic time"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The reported 'shortest possible' times are taken from a multistart local search that has no certificate of global optimality, so if better control fields exist at shorter times the quadratic scalings become upper bounds rather than true minimal times.","fun_headline_variants_meta":{"raw":{"variants":["Single control, quadratic time to entangled states","One knob steers Heisenberg chain to Dicke states in ~N^2 time","Dicke states with a single local pulse: time grows as N^2","Subspace control: reach Dicke states in quadratic time with one qubit","From product to Dicke: one local Z, quadratic time"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000257,"raw_usage":{"total_tokens":1504,"prompt_tokens":923,"completion_tokens":581,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":667,"completion_tokens_details":{"reasoning_tokens":502}},"tokens_in":667,"tokens_out":581,"duration_ms":5280,"temperature":1.0,"reasoning_tokens":502,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T13:21:04.145477+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For any N in 3–9, run a more exhaustive search (e.g., M ≥ 30 harmonics, multiple global optimizers, or a grid of T below the reported T_min) and check whether a control field with the same fidelity threshold 1−10^-3 exists. If one is found, the claim of shortest-possible time fails. Alternatively, an analytic lower bound on T_min growing faster than N^2 would falsify the extrapolated scaling.","supporting_citations":[],"review_version":1}