{"id":"945d70ca-dd7f-4efd-ac38-f6121a9834d3","arxiv_id":"2601.10559","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A numerically optimized sequence of Jaynes-Cummings pulses and displacements prepares Fock states with fidelity above 0.95 up to about 160 photons and success probability above 0.90 in simulation.","lead":"This paper proposes a control recipe made of Jaynes-Cummings pulses and displacements that reshapes a laser-like state into a precise photon-number state, reaching fidelities above 0.95 in simulations for up to about 160-200 photons. A generalist might care because such Fock states are a practical resource for quantum sensing and bosonic quantum computing, and the recipe uses operations already available in several experimental platforms.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Reported fidelities rest on a single non-convex optimization pipeline; without multi-seed or independent verification, the scaling claim is not yet established.","rationale":"The reader's conditional verdict is appropriate. The paper's central claim is a numerical scaling result, and the numerical pipeline has not been made reproducible. The paper's own Fig. 5 and Appendix B acknowledge a rugged, quasi-periodic landscape with dense local minima, while Appendix C describes a transfer-initialized GAdam routine with no released artifacts. A single optimization trajectory under a fixed budget cannot establish that the reported fidelities are intrinsic to the protocol. This is the most load-bearing concern because every downstream statement — scalability, near-determinism, robustness — inherits from the Fig. 2(a) values. The concrete test of multi-seed re-optimization would settle whether the reported F ≥ 0.95 at N = 160/200 is robust or an artifact. I do not see a more fundamental mathematical inconsistency in the JC+displacement construction; the fidelity metric and the revival-time reasoning are plausible. The verdict should remain CONDITIONAL pending these reproducibility checks.","tokens_in":15045,"tokens_out":19427,"duration_ms":194367,"concrete_test":"Re-run the GAdam optimization for N = 160 and N = 200 with at least 50 independent transfer-initialized seeds and a 5× larger optimization budget; record the best and median post-selected fidelity across seeds. If the maximum over seeds falls below 0.95, or if the seed-to-seed spread exceeds 0.05, the Fig. 2(a) trend is not a robust property of the protocol and the central claim should be weakened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim — post-selected F ≥ 0.95 up to N ≈ 160 and P_succ ≳ 0.90 — is produced entirely by the hybrid GAdam optimizer (Appendix C). The paper itself states (Appendix B, Fig. 5) that the loss landscape is quasi-periodic and dense with local minima. All reported points come from a single transfer-initialized optimization trajectory under a fixed budget, with no released code, no optimized parameter sets, no multi-seed statistics, and no convergence diagnostics. The claim 'the protocol achieves' therefore rests on the optimizer having found near-global minima; if the reported runs are local-minimum artifacts, the observed weak scaling with N would not reflect an intrinsic property of the JC–displacement ansatz. The text's assertion that the fixed budget makes the result 'conservative' is only valid if the search is already close to the global minimum, which is exactly what is unverified. This is a load-bearing reproducibility/correctness concern, not a stylistic one: without an independent or multi-start confirmation, a reader cannot distinguish a genuine scaling law from a favorable optimization trajectory.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a control protocol for preparing large photon-number (Fock) states by alternating Jaynes–Cummings (JC) evolutions with phase-space displacements, starting from a coherent state with mean photon number N and optionally post-selecting on a final qubit measurement. The control parameters (pulse durations, displacements, and projection angles) are optimized with a hybrid genetic–Adam routine. The central numerical claim is that post-selected fidelities F ≥ 0.95 are achieved for all targets up to N = 160, remain slightly below 0.95 at N = 180 and 200, and the post-selection success probability is ≳0.90. Robustness to detuning, pulse errors, and dissipation is also discussed. Appendices derive the revival-time condition, document the rugged loss landscape, and describe the GAdam optimizer.","tokens_in":15346,"tokens_out":3470,"duration_ms":41520,"significance":"If the numerical results are reproducible and correctly interpreted, the protocol would be a meaningful step toward preparing few-hundred-photon Fock states with shallow circuits using native spin–oscillator operations. The revival-time insight and the explicit treatment of the non-convex landscape in Appendix B are useful, and the noise-averaged robustness study in Fig. 3(b) is a positive feature. However, the central scaling claim is entirely a product of a single non-convex optimization pipeline; no optimized parameter sets, no code, no multi-seed statistics, and no convergence diagnostics are provided. The paper also overstates the achieved N-range and the unconditional-fidelity claim. These issues are fixable but are load-bearing for the paper's main conclusion.","major_comments":[{"comment":"The abstract states that the protocol 'already achieves high preparation fidelity unconditionally' and that 'fidelities exceeding 0.95 is achieved for photon numbers in the few-hundred regime.' Both claims are stronger than what Sec. II B and Fig. 2(a) actually show. The reported F values are post-selected fidelities, and they satisfy F ≥ 0.95 only up to N = 160, with N = 180 and 200 'slightly below 0.95.' No unconditional (pre-projection) fidelity data are presented anywhere. Please revise the abstract to match the quantitative results and either add unconditional-fidelity data or explicitly frame the abstract claims as post-selected.","section":"Abstract and Sec. II B (Fig. 2)"},{"comment":"The optimization minimizes the joint-state loss L = 1 − |⟨Ψ(τ,β)|Ψ_N(φ)⟩|², but the reported figure of merit is the reduced-state fidelity F in Eq. (5). These are not the same quantity: the joint overlap equals P_succ × F_cond only after conditioning on the qubit projection, so minimizing L can trade off success probability against conditional fidelity. The paper does not explain how the reported F values are computed from the optimized parameters or how the proxy relates to the reported F. Since the central quantitative claim is the fidelity, this connection must be made explicit and correct.","section":"Eq. (6) and Fig. 2(a)"},{"comment":"The scalability trend in Fig. 2(a) rests entirely on the GAdam optimizer finding sufficiently good local minima. Appendix B explicitly shows a dense, quasi-periodic landscape with many competing minima (Fig. 5), and Appendix C states that results are obtained 'under a fixed optimization budget.' The claim that this makes the result 'conservative' is only valid if the search is close to the global minimum, which is not demonstrated. No multi-seed statistics, independent restart checks, convergence diagnostics, or released parameter sets are provided. Without such verification, a reader cannot distinguish a genuine scaling property of the protocol from a favorable single optimization trajectory. Please add multi-start/independent-verification data and release the optimized parameters or code.","section":"Appendices B and C"},{"comment":"The dissipation robustness conclusion is extrapolated from a cooperativity benchmark rather than from master-equation simulations. The paper introduces a Lindblad equation with κ and Γ, but no open-system simulation results are shown. The estimate F ≳ 0.8 for N = 100 is obtained by combining a unitary-optimized value with a collective cooperativity product N_atom C, which is not a substitute for solving the dissipative dynamics with the actual optimized pulse sequence. Moreover, the collective enhancement Ω_eff = Ω√N_atom is used in a context where multi-photon JC dynamics may not reduce to the single-excitation collective manifold. Please present actual master-equation results for at least one representative N and state the parameter assumptions explicitly.","section":"Sec. II C, Dissipation effects"}],"minor_comments":[{"comment":"Typo: 'Fig. 2 benchmarks the our protocol' should read 'benchmarks our protocol.'","section":"Sec. II B"},{"comment":"Text contains 'and and' in the dissipation paragraph: 'Using an experimentally motivated collective cooperativity benchmark N_atom C ∼ 10^8 and and taking...' Please fix.","section":"Sec. II C"},{"comment":"The definition of c_m is given with α^n in the text, but the index should be m (and n for c_n). The notation s_m' and the m≈n≈|α|² statement would be clearer if the approximation used for s_m were written out explicitly.","section":"Appendix A, Eq. (A4)"},{"comment":"The marker legend for the 'optimized multi-pulse protocol (p≤10)' is not self-explanatory; please indicate which point type corresponds to which N and define the staircase ℓ axis in the caption.","section":"Fig. 2(a)"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses an interesting and timely problem, and the proposed interference-engineering idea is worth considering. However, the central quantitative claims are entirely conditioned on a single non-convex optimization run. In my view, the manuscript requires substantial additional numerical evidence (multi-seed optimization, independent verification, and explicit relation of the loss proxy to the reported fidelity) before it can be accepted. The abstract's overstatement of the N-range and unconditional fidelity should also be corrected. I do not see this as a fundamental flaw in the protocol idea, so revision rather than rejection seems appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read the paper. The idea is simple and physically sensible: instead of one long JC interaction at revival time, use many short JC pulses interleaved with displacements. That gives you much more interference control, and the numbers are surprisingly good — post-selected fidelities above 0.95 out to N = 160, with success probability > 0.9. If that trend is real, it is a practical advance for Fock-state generation in cavity QED and trapped ions. The paper is also honest: it shows the loss landscape is rugged, describes the optimizer in detail, and discloses the overlapping Kerr-based preprint in a note added.\n\nThe main problem is that the scaling trend is produced by one optimizer run. The paper says the landscape is quasi-periodic and full of local minima; every point in Fig. 2(a) comes from a single GAdam run with transfer initialization. There is no multi-seed statistics, no released code, no parameter sets, and no convergence diagnostics. The authors say the results are conservative because of a fixed budget, but a fixed budget only makes results conservative if you are already near the global minimum — which is exactly what is not shown. This is a load-bearing reproducibility issue, not a style issue.\n\nTwo smaller overstatements. The abstract says F ≥ 0.95 for photon numbers in the few-hundred regime, but the plot shows that at N = 180, 200 the fidelity dips slightly below 0.95. And the abstract says the protocol already achieves high fidelity unconditionally, but no unconditional numbers are shown — all the quantitative results are post-selected. The dissipation robustness is also extrapolated from a cooperativity benchmark rather than simulated; the detuning and pulse-error plots are real simulations, but the master equation is never integrated for the optimized sequences.\n\nNone of this kills the paper. The physics is standard, the optimization strategy is thoughtful, and the protocol itself is a legitimate extension of Ref. [38] with different ingredients than the Kerr-based approach. For a reader working on bosonic state preparation, the paper is worth knowing about. I would send it to peer review, but a competent referee should ask for multi-seed optimization runs (or at least convergence diagnostics), a plot of unconditional fidelity, and a master-equation simulation at one representative N with realistic rates. With those additions, the central claim would be much more convincing.","headline":"Plausible and useful numerical protocol for Fock-state preparation, but the headline fidelities rest on a single optimizer run and the dissipation claim is extrapolated, so the scaling needs independent verification.","tokens_in":15854,"tokens_out":3740,"would_cite":true,"duration_ms":40074,"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 sequence of alternating Jaynes–Cummings pulses and displacements can prepare Fock states with N up to about 200 at post-selected fidelity above 0.95 and success probability above 0.90, using shallow control circuits.","keywords":["Fock states","large photon-number states","Jaynes–Cummings model","coherent-state interference","quantum control","cavity QED","variational pulse optimization","near-deterministic state preparation"],"falsifier":"Reoptimize the pulse sequences for N=160, 180, and 200 from many independent, non-transfer random initializations, or with a substantially larger optimization budget, and compare the best and median fidelities with the paper's reported values; if the best fidelity drops below 0.95 or scatters widely across seeds, the claimed scaling is not robust. Conversely, an experiment on a cavity-QED or circuit-QED platform that implements the optimized sequence and reconstructs the Wigner function or photon-number distribution at N≈100 should show a dominant peak at n=N with fidelity near the reported le","tokens_in":14919,"feed_emoji":"⚛️","tokens_out":5916,"duration_ms":63134,"temperature":0.7,"pith_summary":"The paper claims that large photon-number (Fock) states can be prepared almost deterministically by interleaving resonant Jaynes–Cummings evolutions with phase-space displacements. The nonlinear photon-number-dependent phases accumulated during the spin–oscillator interaction are converted, by the displacements, into constructive interference that funnels a coherent seed state into a sharp photon-number peak. With optional final qubit projection, post-selected fidelities stay at or above 0.95 up to N=160 and only slightly below at N=180 and 200, with success probability above 0.90. The optimized sequences remain shallow, at most ten layers, and are shown to tolerate detuning, control noise, and dissipation at levels relevant for current platforms. If these numerical results hold, the protocol offers a concrete route to non-Gaussian states that previously required more resource-intensive or probabilistic methods.","feed_headline":"Pulse recipe reaches 200-photon Fock states at high fidelity","feed_subtitle":"Jaynes–Cummings interactions plus displacements give near-deterministic, shallow-depth generation of large photon-number states.","key_machinery":"The central object is the composite pulse: one resonant Jaynes–Cummings evolution segment U(τ_k)=exp(−iHτ_k), governed by H=−Δa†a + Ω(aσ_+ + a†σ_−), followed by a phase-space displacement D(β_k). Repeating these layers p times, with the total time structured by the revival index ℓ through Ω T_R^(ℓ)≈(2ℓ+1)π/√N, imprints photon-number-dependent phases and converts them, via displacement-induced mixing, into constructive interference in Fock space. An optional final qubit projection removes residual qubit–field correlations and enhances the cavity-state purity.","core_discovery":"On the paper's own terms, the central discovery is that a control sequence built from native spin–oscillator operations—alternating JC interaction segments U(τ_k) with displacement pulses D(β_k), starting from an excited qubit and a coherent field with mean photon number |α|²=N—can reshape that coherent state into a Fock state at large N. The optimized multi-pulse sequences (p≤10) achieve post-selected fidelity F≥0.95 for all targets up to N=160, with F only slightly below 0.95 at N=180 and 200, and success probability P_succ≳0.90 for the final qubit projection. The dynamics work by imprinting number-dependent phases through the nonlinear JC spectrum and then using displacements to convert t","pith_inferences":["Editorial inference: the reported scaling trend may be influenced by the fixed optimization budget on a rugged quasi-periodic landscape; a natural check is to compare transfer-initialized searches with many random restarts at N≈200 under a much larger budget. If random restarts match the transfer-seeded results, the fidelity trend is a property of the protocol; if not, it is partly an optimizer pr","Editorial inference: the saturation of detuning tolerance with total evolution time, rather than with pulse segmentation, implies that increasing the coupling strength (for example through collective coupling of many emitters) should simultaneously shorten the sequence and widen the usable detuning window. The paper's dissipation estimate leans on collective enhancement, but the detuning-side bene","Editorial inference: because each JC segment redistributes population only locally while displacements perform the long-range envelope reshaping, one could combine this protocol with a weak continuous measurement or a second photon-number post-selection to trade a small success probability for additional fidelity gain; the paper does not discuss such hybrid strategies.","Editorial inference: the phase-texture alignment near n+m≈2N suggests that stopping one step earlier or changing the final displacement and projection should produce Fock-state superpositions rather than only pure Fock states. The paper mentions superpositions in outlook but does not quantify the achievable fidelity for them."],"forward_implications":["If correct, the protocol removes the need for Kerr-type nonlinearities or measurement-based filtering when generating large Fock states: only linear displacements and the intrinsic Jaynes–Cummings interaction are required.","Because the total evolution time is set by ΩT≈(2ℓ+1)π/√N, the time cost grows only polynomially with N rather than linearly in the number of added photons, and the pulse count p grows slowly, staying at most ten over the explored range.","Post-selected preparation remains near-deterministic, with success probability above 0.90, so the final qubit measurement does not reintroduce the exponential overhead typical of heralded Fock-state sources.","The demonstrated robustness to sub-percent timing and displacement errors, as well as to detuning, suggests the sequences are compatible with existing cavity-QED, circuit-QED, and trapped-ion platforms without demanding error correction.","The same interference-engineering language is intended to extend to other non-Gaussian states, such as photon-number superpositions and grid-like states, as the paper states in its outlook."],"fun_headline_variants":["Near-deterministic Fock states up to 200 photons","High-fidelity 200-photon Fock states via spin-oscillator control","Scalable Fock-state prep with >0.9 success probability","Spin-oscillator pulses yield large Fock states with >95% fidelity","200-photon Fock states made near-deterministically"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The result rests on the numerical assumption that the hybrid genetic-Adam search, run under a fixed optimization budget in a quasi-periodic loss landscape full of local minima, finds near-global optima; if it gets trapped, the reported 'only weak degradation with N' may not reflect the protocol's true capability.","fun_headline_variants_meta":{"raw":{"variants":["Near-deterministic Fock states up to 200 photons","High-fidelity 200-photon Fock states via spin-oscillator control","Scalable Fock-state prep with >0.9 success probability","Spin-oscillator pulses yield large Fock states with >95% fidelity","200-photon Fock states made near-deterministically"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001163,"raw_usage":{"total_tokens":4675,"prompt_tokens":790,"completion_tokens":3885,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":534,"completion_tokens_details":{"reasoning_tokens":3792}},"tokens_in":534,"tokens_out":3885,"duration_ms":27044,"temperature":1.0,"reasoning_tokens":3792,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T10:14:24.663000+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Reoptimize the pulse sequences for N=160, 180, and 200 from many independent, non-transfer random initializations, or with a substantially larger optimization budget, and compare the best and median fidelities with the paper's reported values; if the best fidelity drops below 0.95 or scatters widely across seeds, the claimed scaling is not robust. Conversely, an experiment on a cavity-QED or circuit-QED platform that implements the optimized sequence and reconstructs the Wigner function or photon-number distribution at N≈100 should show a dominant peak at n=N with fidelity near the reported le","supporting_citations":[],"review_version":1}