{"id":"e6bfef91-2758-44e1-bdf5-c3144f82f2a7","arxiv_id":"2502.06465","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A Rydberg-blockaded atomic array can act as a controllable qudit whose dimension grows with atom number, with explicit pulse sequences for arbitrary quantum gates.","lead":"This paper proposes using a Rydberg-blockaded array of atoms as a multi-level quantum memory, or qudit, controlled by global laser pulses. It gives explicit pulse sequences for arbitrary state preparation and quantum gates, with numerical estimates of the resulting errors.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The printed resonance condition in Sec. III.b (Δ01 = ±(Ω1r/2)(√(q+1)−√q)) does not make the intended dressed-state pair degenerate; the correct detuning is ±(Ω1r/2)(√(q+1)+√q), so the pulse sequence as written is internally inconsistent.","rationale":"The geometric feasibility limitation identified by the reader is a quantitative constraint: the hard-blockade assumption may restrict the maximum N for a given atomic spacing and Rydberg state, but it does not invalidate the protocol for the N values demonstrated numerically (N=7,8). The detuning inconsistency, however, directly undermines the derivation of the central control mechanism: if the printed formula were used, the system would not be resonantly driven in the intended subspace, and the 'fold' rotations would not map the state as claimed. This is an internal mathematical error, not a mere disagreement with experimental consensus. Because the paper provides no code or raw data, the reader cannot tell whether the numerical results were obtained with the printed formula or a corrected one; as submitted, the protocol is not reproducible. This fully supports the reader's CONDITIONAL verdict: the underlying idea is likely sound, but the manuscript requires a concrete correction (the plus sign in the detuning) and ideally an explicit release of the simulation code. The geometric feasibility should also be checked quantitatively for the N=200 state-preparation claim, but that is a secondary concern.","tokens_in":12006,"tokens_out":28384,"duration_ms":213197,"concrete_test":"Re-derive the resonance condition from Eq. (1) and the detuning term; then implement the Sec. III.b pulse sequence with the printed detuning Δ01 = ±(Ω1r/2)(√(q+1)−√q) and with the corrected sum ±(Ω1r/2)(√(q+1)+√q) for N=7, and compare the resulting gate infidelity to Fig. 4(c). If the printed formula fails while the corrected one reproduces the reported ~N^3(Ω01/Ω1r)^2 scaling, the error is a typographical one that must be fixed before the protocol can be reproduced.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Sec. III.b ('Full control over the Hilbert space'), the protocol isolates the two-level subspace {|±,q+1⟩,|∓,q⟩} by setting Δ01 = ±(Ω1r/2)(√(q+1)−√q) and ϕ1r=0. This is not the resonance condition. From Eq. (1) and the detuning term −Δ01 Σ_j(|1_j⟩⟨1_j|+|r_j⟩⟨r_j|), which contributes −Δ01 q to every |±,q⟩, the energy difference between |+,q+1⟩ and |−,q⟩ is Ω1r(√(q+1)+√q)/2 − Δ01; resonance requires Δ01 = +Ω1r(√(q+1)+√q)/2, and for |−,q+1⟩↔|+,q⟩, Δ01 = −Ω1r(√(q+1)+√q)/2. With the printed difference formula, the two states are off-resonant by Ω1r√q (or Ω1r√(q+1)), so the effective Hamiltonian H_eff written below Eq. (2) is not the actual dynamics: the intended rotation does not occur, and the 'fold' operation is invalid. Since no code or data are provided, the numerical gate fidelities (Figs. 3–4) cannot be checked against the manuscript's equations; the text is not self-consistent, and the central protocol is not reproducible as written.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a protocol for universal qudit control using the collective dressed states of a Rydberg-blockaded array of N three-level atoms, which are isomorphic to the Jaynes-Cummings ladder. The protocol uses global laser pulses to prepare arbitrary states and implement arbitrary unitaries on the 2N-dimensional qudit Hilbert space, with numerical demonstrations for N=7 (14-level qudit) and scaling estimates for gate infidelities and Rydberg-state decay.","tokens_in":12305,"tokens_out":14114,"duration_ms":110014,"significance":"If correct, the protocol would offer a scalable qudit platform with global addressing, avoiding the need for site-selective control, and it builds naturally on established Jaynes-Cummings physics in Rydberg ensembles. The paper provides a clear mapping, explicit pulse constructions based on a known unitary decomposition, numerical simulations for small N, and quantitative estimates of errors and decay. These are genuine strengths. However, the central resonance condition in Sec. III.b is incorrect as written, which currently undermines the main construction. With the necessary corrections and reproducible numerical details, the work could be a useful contribution to Rydberg-based qudit quantum processing.","major_comments":[{"comment":"The resonance condition for the fold operation is stated as Δ01 = ±(Ω1r/2)(√(q+1)−√q). This is incorrect. Using Eq. (1) and the detuning term −Δ01 Σ_j(|1_j⟩⟨1_j|+|r_j⟩⟨r_j|) of H_c, the energy difference between |+,q+1⟩ and |−,q⟩ is Ω1r(√(q+1)+√q)/2 − Δ01, so the resonance condition for this pair is Δ01 = +Ω1r(√(q+1)+√q)/2; for the opposite-symmetry pair it is Δ01 = −Ω1r(√(q+1)+√q)/2. With the printed minus sign, the selected states are off-resonant by Ω1r√q or Ω1r√(q+1), so the effective Hamiltonian H_eff written below Eq. (2) does not describe the actual dynamics, and the intended fold rotation is not realized. This invalidates the central construction of Sec. III.b unless corrected; the numerical simulations of Sec. V must be checked against the corrected condition.","section":"Sec. III.b"},{"comment":"The manuscript does not provide the numerical parameters used in the simulations (e.g., the Rabi frequencies, detunings, and pulse durations for the examples in Figs. 3 and 4), nor the code or an explicit algorithm to generate the pulse sequences. Since the pulse sequence depends on the target state through computed rotation angles, this lack of detail prevents reproduction of the reported gate fidelities. In particular, with the resonance condition error of Comment 1, it is impossible to verify that the simulations correspond to the printed protocol.","section":"Secs. V and VI"},{"comment":"The large-N feasibility estimates, e.g., 400-level state preparation with N=200 atoms, do not check the geometric constraints N^(1/d) a < R_b and a > λ simultaneously. For a two-dimensional array with a > λ ≈ 1 μm, N=200 requires a side length exceeding 14 μm, and hence a blockade radius R_b > 14 μm; the authors should verify that this is compatible with the chosen Ω1r and the atomic C6 coefficient, or else the scalability claim is not supported.","section":"Sec. VII"}],"minor_comments":[{"comment":"The word 'Rydbgerg' in the abstract should be 'Rydberg'.","section":"Abstract"},{"comment":"The caption refers to panel '(d)' but the figure contains only panels (a)–(c); the reference should likely be to (c).","section":"Fig. 2 caption"},{"comment":"The text states the initial state is |ψ_in⟩ = |ψ_target⟩, while the caption of Fig. 2 uses |ψ_in⟩ = e^{−iπ/4}|ψ_target⟩; clarify this inconsistency.","section":"Sec. V.a"},{"comment":"The expression for |ψ_after⟩ after the fold rotation contains a normalization factor that appears to be a typo; it should be written as sqrt(|a_{∓,q}|^2 + |a_{±,q+1}|^2) times |∓,q⟩, not the printed form.","section":"Sec. III.b"},{"comment":"The infidelity metric ϵ = 1 − (1/(2N)^2)|Tr(U_target† U)|^2 is the process infidelity, but the text does not define it as such; define it clearly and consider relating it to the average gate fidelity.","section":"Sec. VI"},{"comment":"The two-pulse echo eR that cancels phase accumulation deserves a more explicit justification; the cancellation relies on the spectra at ϕ1r=0 and ϕ1r=π being opposite, which is true but not stated.","section":"Sec. III.b"}],"recommendation":"major_revision","confidential_remarks":"The resonance condition error appears to be a sign typo, but it affects the core protocol and must be fixed. The paper's central idea is plausible and the numerical results may be valid if the simulations used the corrected condition, but the authors need to supply reproducible parameters or code. The novelty is incremental relative to existing Rydberg-ensemble control work, but the global-addressing qudit scheme is a reasonable contribution if the technical issues are resolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core idea is sound: encode a qudit in the collective dressed states of a Rydberg blockaded array and control it with globally addressed pulses. The paper gives explicit pulse sequences for arbitrary state preparation and unitaries, and the numerical simulations for N=7 (14-level qudit) show gate fidelities that are consistent with the claimed scaling. The approach avoids site-selective addressing, which is a real practical advantage. That part deserves credit.\n\nThe main problem is in Sec. III.b. The paper sets Delta_01 = ±(Omega_1r/2)(sqrt(q+1)-sqrt(q)) to isolate the subspace {|±,q+1>, |∓,q>}. That is not resonant. The energy difference between |+,q+1> and |-,q> is Omega_1r(sqrt(q+1)+sqrt(q))/2, and the detuning term shifts |±,q> by -Delta_01 q, so the correct resonance condition is Delta_01 = ±(Omega_1r/2)(sqrt(q+1)+sqrt(q)). With the printed difference formula, the pair is off-resonant by Omega_1r sqrt(q) (or sqrt(q+1)), so the effective Hamiltonian written below Eq. (2) does not describe the actual dynamics. The protocol as written is not reproducible. I suspect it is a typo—the numerics presumably used the correct sum—but the text is internally inconsistent, and no code or data are provided to check.\n\nA secondary soft spot: the geometric constraints for large qudits are not checked. The paper claims state preparation with N=200 (400-level qudit) but doesn't verify that N^(1/d) a < R_b and a > lambda can both be satisfied for those parameters. For the experimentally relevant range (N up to 8-14 atoms for complex gates), this is probably fine, but the scaling claims need a quantitative feasibility check.\n\nThe scaling analysis for gate infidelity and the Rydberg decay budget are sensible, and the conclusion that complex gates are limited to about 16-level qudits is honest. The novelty relative to Mischuck-Mølmer and Keating is modest but real: the explicit, globally addressed pulse sequence and the error scaling for a Rydberg blockaded array are not in the prior work.\n\nThis deserves peer review. The error is likely fixable, and the underlying protocol is useful for the neutral-atom qudit community. A careful referee should ask for the corrected detuning, ideally accompanied by simulation code or data to confirm the numerics match the equations. I would not cite the paper in its current form, but I would read a revised version.","headline":"A useful qudit-control protocol for Rydberg blockaded arrays, but the printed resonance condition for the folding rotations is a sign error that makes the central pulse sequence inconsistent as written.","tokens_in":12869,"tokens_out":5559,"would_cite":false,"duration_ms":43352,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.Lx","32.80.Ee"],"model":"deepseek-v4-flash","headline":"A Rydberg-blockaded array of N three-level atoms can encode a fully controllable qudit of dimension 2N, with arbitrary state preparation and arbitrary unitaries achieved by global laser pulse sequences.","keywords":["qudit","Rydberg blockade","dressed states","Jaynes-Cummings model","atomic arrays","pulse sequences","state synthesis","unitary gates"],"falsifier":"Measure the gate infidelity of the generalized Hadamard gate for N=8 atoms as a function of Ω01/Ω1r; the paper predicts ϵ ≈ $N^{3}$ (Ω01/Ω1r)^2, so a measured scaling significantly worse than this, or clear population leakage out of the symmetric single-excitation subspace seen in spectroscopic resolution of |±,q⟩, would settle against the central claim.","tokens_in":11774,"feed_emoji":"⚛️","tokens_out":5012,"duration_ms":42993,"temperature":0.7,"pith_summary":"The paper proposes using a Rydberg-blockaded array of N identical three-level atoms as a single qudit whose Hilbert space is the 2N collective dressed states of the Jaynes–Cummings ladder. It shows that a second, weaker laser can implement any rotation among these dressed states, so any target state can be synthesized and any unitary can be approximated by concatenating generalized phase gates. The qudit dimension grows simply by adding atoms, and the whole protocol needs only global laser pulses, no site-selective addressing. Numerical simulations for N=7 confirm gate infidelities scaling as $N^{3}$ (Ω01/Ω1r)^2, and estimates including Rydberg decay place practical limits near a 400-level qudit for state preparation.","feed_headline":"A Rydberg atom array becomes a large controllable qudit","feed_subtitle":"Global laser pulses steer a 2N-level qudit; adding atoms scales the code space.","key_machinery":"The key object is the Jaynes–Cummings ladder of collective dressed states |±,q⟩ = (|e,q−1⟩ ± |g,q⟩)/√2, where q counts atoms in the intermediate state. The control-laser Hamiltonian projected onto this ladder produces couplings with prefactors K_N^q and Q_N^q that allow rotations in two-dimensional subspaces; pulse sequences built from these effective two-level rotations implement full control over the qudit Hilbert space and the generalized phase gate on |−,1⟩.","core_discovery":"The central discovery is that the collective Hilbert space of a hard-blockaded array—one Rydberg excitation shared among N atoms—is a realization of the Jaynes–Cummings model whose dressed states |±,q⟩ form a 2N-dimensional qudit. Because the dressed-state energies depend nonlinearly on q through ±(Ω1r/2)√q, the detuning and phase of the laser driving the intermediate-to-Rydberg transition act as knobs that set the level structure, while the laser driving the ground-to-intermediate transition couples adjacent rungs. The authors show these couplings can be used to fold any state to a reference state |−,1⟩, apply a phase gate there, and unfold back, giving arbitrary unitaries; numerical integration of the full (not just effective) Hamiltonian demonstrates the gates.","pith_inferences":["Beyond the paper's estimates, the N^3 pulse count means that control-field amplitude and phase noise, which are not modeled here, may set the practical fidelity floor for large N before Rydberg decay does.","Because the encoding relies on exact collective symmetry, any atom loss or position disorder during a sequence changes the effective qudit dimension and invalidates the pulse sequence; the paper assumes defect-free arrays but does not analyze robustness to loss.","The same dressed-state ladder could be repurposed for bosonic-like error-correcting codes; the paper's outlook mentions permutation-invariant codes, so a natural next step would be to test deletion-error correction on this qudit platform."],"forward_implications":["The qudit dimension is set by atom number, so scaling the qudit becomes an experimental loading problem rather than a redesign of the control architecture.","Any target unitary can be built from O(N^2) pulses via eigen-decomposition into generalized phase gates, and the pulse sequence can be computed ab initio.","Generalized phase gates on a 14-level qudit reach infidelity about 9×10^-5 at Ω01/Ω1r = 10^-3; infidelity scales as N^2 for phase gates and N^3 for arbitrary gates.","Rydberg-state decay bounds the useful size: the paper estimates up to a 16-level qudit for a generalized Hadamard gate, 28-level for phase gates, and 400-level for state preparation under cited experimental parameters.","Because only global addressing is required, the protocol avoids site-selective single-atom addressing."],"supporting_citations":[{"why":"Supplies the Jaynes–Cummings model whose ladder spectrum the dressed states reproduce.","marker":"[16]"},{"why":"Establishes the mapping of a Rydberg blockaded ensemble to the spin-boson/Jaynes–Cummings Hamiltonian and demonstrates arbitrary Dicke-state control.","marker":"[18]"},{"why":"Gives the experimental demonstration of the Jaynes–Cummings ladder with Rydberg-dressed atoms, underpinning the spectroscopic measurement scheme.","marker":"[19]"},{"why":"Provides the earlier proposal for qudit quantum computation in the Jaynes–Cummings model that this protocol extends.","marker":"[20]"},{"why":"Supplies the decomposition of arbitrary multivalued logic gates into generalized phase gates on target vectors, the method used to build unitaries.","marker":"[24]"},{"why":"Provides the state-of-the-art Rabi frequency value Ω1r/(2π)=300 MHz used in the fidelity and decay estimates.","marker":"[25]"},{"why":"Provides the Rydberg-state lifetime calculations used to estimate relaxation probabilities for the pulse sequences.","marker":"[26]"}],"fun_headline_variants":["Rydberg blockade array yields a 2N-level qudit","Qudit from a Jaynes-Cummings Rydberg superatom","Arbitrary unitaries on a Rydberg blockaded atom qudit","2N-level qudit from collective Rydberg states","Pulse-controlled Rydberg superatom acts as scalable qudit"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire protocol assumes all N atoms are inside the blockade radius, are addressed identically by both lasers, and remain in the symmetric subspace with at most one Rydberg excitation throughout; if atom positions or atom number fluctuate, the dressed-state ladder and the pulse sequence no longer describe the system.","fun_headline_variants_meta":{"raw":{"variants":["Rydberg blockade array yields a 2N-level qudit","Qudit from a Jaynes-Cummings Rydberg superatom","Arbitrary unitaries on a Rydberg blockaded atom qudit","2N-level qudit from collective Rydberg states","Pulse-controlled Rydberg superatom acts as scalable qudit"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000958,"raw_usage":{"total_tokens":4039,"prompt_tokens":862,"completion_tokens":3177,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":478,"completion_tokens_details":{"reasoning_tokens":3084}},"tokens_in":478,"tokens_out":3177,"duration_ms":20215,"temperature":1.0,"reasoning_tokens":3084,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T15:22:10.482040+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the gate infidelity of the generalized Hadamard gate for N=8 atoms as a function of Ω01/Ω1r; the paper predicts ϵ ≈ $N^{3}$ (Ω01/Ω1r)^2, so a measured scaling significantly worse than this, or clear population leakage out of the symmetric single-excitation subspace seen in spectroscopic resolution of |±,q⟩, would settle against the central claim.","supporting_citations":[{"cited_title":"Keating, C","cited_arxiv_id":null,"evidence_quote":"Establishes the mapping of a Rydberg blockaded ensemble to the spin-boson/Jaynes–Cummings Hamiltonian and demonstrates arbitrary Dicke-state control."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the experimental demonstration of the Jaynes–Cummings ladder with Rydberg-dressed atoms, underpinning the spectroscopic measurement scheme."},{"cited_title":"Mischuck and K","cited_arxiv_id":null,"evidence_quote":"Provides the earlier proposal for qudit quantum computation in the Jaynes–Cummings model that this protocol extends."},{"cited_title":"Muthukrishnan and C","cited_arxiv_id":null,"evidence_quote":"Supplies the decomposition of arbitrary multivalued logic gates into generalized phase gates on target vectors, the method used to build unitaries."},{"cited_title":"4 (c) the regions where the relaxation probability Prelax is smaller that the gate infidelity ϵ","cited_arxiv_id":null,"evidence_quote":"Provides the state-of-the-art Rabi frequency value Ω1r/(2π)=300 MHz used in the fidelity and decay estimates."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Rydberg-state lifetime calculations used to estimate relaxation probabilities for the pulse sequences."}],"review_version":1}