{"id":"492f37cc-7971-4b66-a715-b2f69a69b0dd","arxiv_id":"1908.04831","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A variational coherent-state method applied after a specially chosen unitary transformation reproduces beyond-mean-field quantum dynamics, including collapse of coherence, at near-mean-field numerical cost.","lead":"This paper introduces a new approximation, the scrambled mean field method, for simulating quantum dynamics of trapped, weakly interacting Bose gases beyond the Gross-Pitaevskii equation. It moves interaction-induced energy shifts into the unperturbed part of the Hamiltonian before applying a coherent-state ansatz, giving equations nearly as simple as mean field while capturing collapse of coherence.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Fig. 2 relies on uncontrolled g t approximations that are not checked against the exact Eq. (7) and likely exceed the stated validity range.","rationale":"I read the paper in good faith: the method is clearly presented, the two-mode benchmark against exact diagonalization with energy conservation to about 1% is genuine evidence for the exact Eq. (6) in the two-mode sector, and the author openly states in the final paragraph that scattering into initially empty modes is beyond the method's scope. That stated limitation is real but does not by itself invalidate the narrower claim about dephasing-driven collapse of coherence. The most load-bearing problem is instead in the multimode implementation. The exact Eq. (7) is replaced, for numerical efficiency, by approximations that discard O(g t) phases; the paper even gives the validity condition g_{jj'}t ≲ 1 and, in the Supplemental, t ≲ hbar/Uc. The reported Josephson simulation runs to times where Uc t/hbar reaches 5 (with hbarω_J = 30Uc), and no mode-by-mode check of g_{jj'}t is provided. Since the 21-mode calculation is the only evidence that the method works beyond two modes, this uncontrolled approximation is the critical weak point. A reduced-mode exact-versus-approximate comparison would settle whether the predicted damping and phase variance are genuine. I therefore keep the reader's CONDITIONAL verdict, with the added condition that the approximate multimode solver be validated against the exact Eq. (7) for a small mode set.","tokens_in":9551,"tokens_out":31361,"duration_ms":311049,"concrete_test":"Solve the Josephson junction with a reduced M=7 mode set using the unapproximated Eq. (7), including the full ν from Eq. (8) and the exact exponential factors, and compare the resulting ⟨φ̄⟩(t) and Var{φ̄}(t) with the Gaussian-sampled approximate solver used for Fig. 2. If the final variance or the damping rate differs by more than about 20% at ω_J t = 150, the Fig. 2 demonstration does not support the central claim. Also report max_{j,j'} |g_{jj'}| t over the initially occupied modes to check whether the stated g t ≲ 1 condition is actually satisfied.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The clean two-mode benchmark evaluates Eq. (6) exactly, but the 21-mode Josephson application implements Eq. (7) through two uncontrolled simplifications: dropping the g-dependent pair phases in ν (Eq. 8) by neglecting non-commutativity of b_j and A_{j'}, and replacing exp[-Σ|ψ|^2(1-e^{iGt})] by a small-time Gaussian expansion. Both are perturbative in g_{jj'}t, yet the collapse of coherence that the method is designed to capture occurs precisely at g_{jj'}t ~ 1, where the dropped terms are order one. The Supplemental states the simplifications are valid for t ≲ hbar/Uc, while Fig. 2 runs to ω_J t = 150 with hbarω_J = 30Uc, i.e., Uc t/hbar = 5 at the end, exceeding the stated range unless all occupied couplings are far below Uc/hbar, which is not shown. The approximate solver is never compared with the unapproximated Eq. (7), even for a small mode number; the reported damping of phase oscillations and growth of phase variance may therefore be artifacts of the factorization rather than of the scrambled mean-field method. This is load-bearing because Fig. 2 is the paper's only multimode demonstration of the central claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a 'scrambled mean field' method for the quantum dynamics of trapped, weakly interacting Bose gases. The key idea is to transform the many-body Hamiltonian to an interaction representation with respect to the harmonic part plus the diagonal quartic part, and then apply a product-of-coherent-states variational ansatz in that representation. The resulting equations, Eq. (6) with Eq. (7), are claimed to be almost as simple as the Gross-Pitaevskii equation while capturing non-mean-field effects such as the collapse of coherence. The method is benchmarked against exact diagonalization for a two-dimensional anharmonic oscillator (Fig. 1 and Supplemental Material), with reported energy conservation to about 1%. It is then applied to a 21-mode bosonic Josephson junction, where it predicts damping of phase oscillations and growth of phase variance (Fig. 2). The final paragraph states that scattering of quanta into initially empty modes remains beyond the method's scope.","tokens_in":9820,"tokens_out":5535,"duration_ms":58144,"significance":"If the central claim is valid, the method offers a computationally cheap way to describe certain multimode quantum-dephasing phenomena in finite-size bosonic systems, a regime where truncated Wigner methods lose accuracy and MCTDHB may require impractically many configurations. The 2D benchmark is encouraging: the method reproduces expectation values, variances, and covariance for a two-mode anharmonic system over long times, with good energy conservation. The proposed numerical complexity reduction is also attractive for qualitative long-time estimates. However, the multimode demonstration is not yet convincing: the Josephson-junction simulation relies on uncontrolled small-time approximations that are never checked against the unapproximated Eq. (7), and the validity range stated in the Supplemental is violated by the simulation times in Fig. 2. The paper therefore establishes the potential of the idea but not yet its multimode utility.","major_comments":[{"comment":"The numerical implementation of Eq. (7) uses two uncontrolled simplifications: neglecting the non-commutativity of b_j and A_{j'} (dropping g-dependent terms in the phase ν) and replacing the exponential dephasing factor exp[-|ψ_j|^2(1-e^{iGt})] by a Gaussian small-time expansion. Both are perturbative in g_{jj'}t, but the collapse of coherence that the method is intended to capture occurs at g_{jj'}t ~ 1. The Supplemental states these simplifications are valid for t ≲ ħ/Uc, whereas Fig. 2 runs to ω_J t = 150 with ħω_J = 30Uc, i.e., Uc t/ħ = 5 at the end. Unless all occupied coupling constants are far below Uc/ħ, which is not shown, the run exceeds the stated range. The approximate solver is never compared with the unapproximated Eq. (7), even for a small number of modes, so the reported damping and variance growth may not reflect the actual scrambled mean-field equations. A quantitative comparison, e.g., for M = 3 or 4 modes, is needed to validate the approximation, or the demonstration should be restricted to times within the controlled regime.","section":"§Numerical method, Eq. (7), Fig. 2, Supplemental II"},{"comment":"No convergence check with respect to the number of modes M is presented. The heuristic thermalization argument in the main text estimates a minimal mode number, but the paper does not show that the Josephson-junction results are stable when M is increased or changed (e.g., M = 14, 21, 34). Without such a test, the observed damping and variance growth could be an artifact of truncation rather than a genuine multimode quantum effect. The claim that the method handles a 'nontrivially large number of modes' requires at least a demonstration that the chosen M is sufficient for the observable of interest.","section":"Fig. 2, mode truncation"},{"comment":"The closing paragraph acknowledges that 'the description of scattering of quanta into initially empty modes remains beyond its scope.' This is a severe limitation for the central claim, because scattering into initially empty modes is generic in multimode Bose-gas dynamics, and the paper does not identify conditions under which it can be neglected. In the Josephson-junction application, the initial coherent state with ⟨ρ⟩ = 0 may have empty or nearly empty high-frequency modes; if those modes become populated, the method cannot capture the resulting dynamics. The paper should either justify that such scattering is negligible for the chosen initial state and observables over the simulated times, or qualify the central claim accordingly.","section":"Final paragraph, scope limitation"}],"minor_comments":[{"comment":"The indices κ and λ in g_{κλ} are not explicitly defined in terms of the mode labels in the term V_{j_1...j_l}^{j'_1...j'_{l'}}; please clarify the notation.","section":"Eq. (8)"},{"comment":"The Supplemental states that the 2D benchmark evaluates the r.h.s. of Eq. (6) exactly, but this is not stated clearly in the main text; adding this distinction would help readers understand that Fig. 1 tests the variational ansatz, not the approximate factorizations used in Fig. 2.","section":"Supplemental I"},{"comment":"The stability test against different random vectors σ_γ shows discrepancies of order 10^{-2} rad, but the effect of these fluctuations on the reported variance, not just the mean phase, should be reported, since the variance is the key demonstration of coherence collapse.","section":"Supplemental II"},{"comment":"There are several typographical and formatting issues (e.g., 'Schr¨ odinger', 'we can the find', inconsistent use of hats in Eq. (11)); a careful proofread is recommended.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper presents an interesting and potentially useful variational idea, and the two-mode benchmark is a genuine positive result. However, the only multimode demonstration rests on numerical approximations whose validity is questionable in exactly the regime where the claimed effect occurs, and no convergence or cross-check against the exact equations is provided. I believe the paper can be revised to address this, but in its current form the central multimode claim is not convincingly supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nHere's the short version: the scrambled mean-field idea is real and worth knowing about, but the paper's only multimode demonstration is built on an approximation that the author himself states is valid only for g t ≈ 1, and he runs the simulation to regions where g t is several times larger, without checking against the exact Eq. (7). That means the central claim about collapse of coherence in a multimode system is not yet supported.\n\nWhat's genuinely new: moving the diagonal quartic term into the unperturbed Hamiltonian and then using a coherent-state ansatz in the interaction representation is a simple and clever trick. The derivation is transparent, there are no fitted parameters, and the 2D benchmark against exact diagonalization is clean, with energy conservation at the 1% level. The stochastic averaging scheme for the Gaussian factors is a neat way to avoid O(M^4) tensor explosion. I also credit the author for stating plainly that the method cannot describe scattering into initially empty modes.\n\nThe soft spots, in order of severity. First, the Josephson junction application (Fig. 2) is the only multimode test, and it uses two uncontrolled simplifications: neglecting the non-commutativity of b_j and A_{j'} (which drops order-one phases at g t ~ 1) and replacing the exponential dephasing factor by a small-time Gaussian expansion. The author says these are valid for g_{jj'} t ≈ 1, but the plot runs to ω_J t = 150, and with hbar ω_J = 30 Uc that means Uc t/hbar = 5 at the end. The dephasing that is the whole point of the method becomes strong exactly around g t ~ 1, so the regime where the approximations become uncontrolled is the regime the method is designed to describe. The approximate solver is never compared with the exact Eq. (7), even for a small number of modes. That comparison is essential before the multimode results can be believed. Second, there is no convergence study in the number of modes M, and no code or data. The claim that M=21 is enough is plausible but not demonstrated. Third, the residual interaction is assumed to act as a mean-field drive; the paper's own final paragraph concedes that scattering into empty modes is out of scope, which is a real limitation for many observables.\n\nThe bottom line: the method deserves a serious referee and a revision, but not acceptance in its present form. The 2D test is fine; the multimode test needs to be redone with a check against the unapproximated equations, at least on a smaller system, and the validity range needs to be respected or extended. I would send it to review, but I would flag the Josephson section as load-bearing and the stress-test concern as valid.","headline":"A genuinely new scheme with a clean two-mode benchmark, but the only multimode test relies on uncontrolled approximations that are not validated, so the central claim exceeds the evidence.","tokens_in":10295,"tokens_out":3745,"would_cite":false,"duration_ms":35012,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.75.Kk","03.75.Lm","05.30.Jp","67.85.-d"],"model":"deepseek-v4-flash","headline":"A product of coherent states, applied after a unitary 'scrambling' transformation that diagonalizes the quartic part of the Hamiltonian, suffices to reproduce the quantum collapse of coherence in degenerate Bose gases with equations…","keywords":["Bose-Einstein condensates","beyond mean field","collapse of coherence","coherent states","bosonic Josephson junction","quantum dephasing","interaction representation","Gross-Pitaevskii equation"],"falsifier":"Run the two-mode anharmonic test with quartic coefficients large enough that $g t$ reaches about 1 within the first oscillation period and compare $\\mathrm{Var}\\{x\\}(t)$ from Eq. (6) with exact diagonalization: systematic divergence at those times would show that the neglected commutators and Gaussian factorization set the method's true validity boundary.","tokens_in":9307,"feed_emoji":"⚛️","tokens_out":10739,"duration_ms":106233,"temperature":0.7,"pith_summary":"This paper proposes a way to compute the quantum dynamics of trapped, weakly interacting Bose gases beyond the Gross-Pitaevskii mean-field limit without exponential many-body cost. The trick is to move to the interaction representation with respect to the harmonic Hamiltonian plus its diagonal quartic part, and only then approximate the transformed state by a product of coherent states. The resulting equations of motion are almost as simple as Gross-Pitaevskii, yet occupation-dependent phases from the 'scrambling' step make field coherence decay and phase variances grow. The paper demonstrates the method on a two-dimensional anharmonic oscillator against exact diagonalization and on a 21-mode bosonic Josephson junction, reproducing damping of Josephson oscillations and growth of phase uncertainty.","feed_headline":"Scrambled mean field tracks quantum dephasing in Bose gases","feed_subtitle":"A coherent-state ansatz after a scrambling transform reproduces damping of Josephson oscillations and phase-variance growth in multimode…","key_machinery":"The central object is the scrambling unitary $\\hat U_r(t)=\\exp(-i\\hat H_0 t/\\hbar)$, where $\\hat H_0$ is the harmonic part $\\hat H_h$ plus the diagonal quartic part $\\hat H_{qd}=(\\hbar/2)\\sum_{j,j'} g_{jj'} \\hat b_j^\\dagger \\hat b_{j'}^\\dagger \\hat b_{j'} \\hat b_j$. Under this unitary each annihilation operator becomes $\\hat b_j \\to \\hat A_j \\hat b_j$ with $\\hat A_j = \\exp(-i\\omega_j t - i\\sum_{j'} g_{jj'} \\hat b_{j'}^\\dagger \\hat b_{j'} t)$, producing the occupation-dependent phase factors in the coherent-state expectation value. The ansatz is a product of coherent states for the transformed wavefunction $|\\tilde\\Psi\\rangle$, and the expectation in Eq. (7) is closed-form in the amplitudes $\\psi_j$ and the coupling matrix $g_{jj'}$. For numerical work the paper restricts to times $g_{jj'}t \\lesssim 1$, neglects the non-commutativity of $\\hat b_j$ and $\\hat A_{j'}$, and replaces the Gaussian dephasing factor by an average over pseudorandom parameters $\\sigma$, which reduces the tensor ranks and removes the apparent $O(M^4)$ scaling.","core_discovery":"The paper claims that the scrambled mean-field equations, Eq. (6) together with Eq. (7), capture the essential quantum dephasing of degenerate Bose gases while keeping the numerics at the level of coupled Gross-Pitaevskii-like equations for complex amplitudes $\\psi_j(t)$. Quantum correlations enter through the factor $\\exp[-\\sum_j |\\psi_j|^2(1 - e^{i G t})]$ in Eq. (7), which appears because the scrambling operator $\\hat U_r(t) = \\exp(-i\\hat H_0 t/\\hbar)$ attaches each occupied mode's particle number to every other mode's phase. In the two-mode anharmonic test, the computed expectation values and variances follow the exact diagonalization closely over hundreds of oscillation periods with energy conserved to about 1%; in the multimode Josephson junction, the method yields decaying phase oscillations and growing phase variance rather than the undamped oscillation of the mean-field solution.","pith_inferences":["A testable extension would be to replace the fixed orthogonal pseudorandom vectors in the $\\sigma$-averaging step with fresh noise at every time step; if the dephasing times shift measurably, the Gaussian factorization itself is biasing the results.","The author's stated blind spot, scattering into initially empty modes, implies the method is most reliable for observables carried by initially occupied modes; coupling the scrambled equations to a weak reservoir of empty high-energy modes could extend it toward thermalization, but this goes beyond the paper.","In the Josephson junction example the predicted damping is slower and less complete than the experimental relaxation, which the author attributes to the simplified model; including density-dependent radial widths or coupling to symmetric modes would tell whether the gap is a model limitation or a method limitation.","Since Eq. (7) evaluates the scrambling factors at coherent amplitudes, the method can likely be recast as a stochastic phase-space evolution in the scrambled variables, connecting it to truncated-Wigner or positive-P techniques, though the paper does not make that connection."],"forward_implications":["If the method is right, quantum dephasing in a trapped BEC or quasicondensate is computable with Gross-Pitaevskii-like effort: the same complex amplitudes $\\psi_j(t)$ that specify the mean field also give the collapsed coherence through occupation-weighted phase spreads.","For a multimode bosonic Josephson junction, the method predicts that Josephson oscillations damp and phase variance grows while the integrated interference visibility remains high, a testable signature of multimode quantum dephasing.","The collapse of coherence is governed by the explicit factor $\\exp[-\\sum_{j'} |\\psi_{j'}|^2(1 - e^{i g_{jj'} t})]$, so collapse times can be read off from the occupation-weighted spread of interaction constants $g_{jj'}$, without simulating the many-body Hilbert space.","Because the construction only assumes well-defined elementary excitations, it transfers to any weakly anharmonic bosonic system, including molecular vibrational dynamics, where the example Hamiltonians can be replaced by an appropriate diagonal quartic part.","The method reaches time scales on which configuration-based methods become impractically large, since it evolves only $M$ complex functions and conserves energy to roughly 1-2% in the tested cases."],"supporting_citations":[{"why":"introduces the rotated Hartree approach that this method resembles but simplifies by deriving the transformation from perturbation theory instead of variational parameters.","marker":"[24]"},{"why":"supplies the coherent-state formalism and factorization properties used to evaluate Eq. (7).","marker":"[25]"},{"why":"establishes the limited validity time of the truncated Wigner approximation, motivating a beyond-mean-field method for longer times.","marker":"[5]"},{"why":"reports collapse-and-revival observations in optical lattices that motivate a multimode quantum treatment beyond single-mode theory.","marker":"[16, 17]"},{"why":"provides the measured relaxation to a phase-locked equilibrium in a one-dimensional bosonic Josephson junction against which the model's damped oscillations are compared.","marker":"[29]"},{"why":"describes the configuration-based MCTDHB benchmark whose practical mode count the scrambled mean-field method seeks to bypass.","marker":"[12]"}],"fun_headline_variants":["Scrambled mean field captures quantum dephasing simply","Quantum dephasing from a scrambled Gross-Pitaevskii twist","Scrambling transform puts quantum dephasing in mean-field reach","Bose gas dephasing captured by scrambled mean-field equations","Beyond Gross-Pitaevskii: a scrambling trick reveals quantum coherence loss"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on the premise that after the scrambling transformation a single product of coherent states still describes the system, with the residual interaction acting only as a mean-field drive; the paper itself notes this leaves out scattering into initially empty modes, and the numerics additionally assume times short compared with the interaction-induced phase shifts.","fun_headline_variants_meta":{"raw":{"variants":["Scrambled mean field captures quantum dephasing simply","Quantum dephasing from a scrambled Gross-Pitaevskii twist","Scrambling transform puts quantum dephasing in mean-field reach","Bose gas dephasing captured by scrambled mean-field equations","Beyond Gross-Pitaevskii: a scrambling trick reveals quantum coherence loss"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000252,"raw_usage":{"total_tokens":1490,"prompt_tokens":804,"completion_tokens":686,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":420,"completion_tokens_details":{"reasoning_tokens":595}},"tokens_in":420,"tokens_out":686,"duration_ms":6468,"temperature":1.0,"reasoning_tokens":595,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:32:04.970029+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the two-mode anharmonic test with quartic coefficients large enough that $g t$ reaches about 1 within the first oscillation period and compare $\\mathrm{Var}\\{x\\}(t)$ from Eq. (6) with exact diagonalization: systematic divergence at those times would show that the neglected commutators and Gaussian factorization set the method's true validity boundary.","supporting_citations":[{"cited_title":"Kucar, H.-D","cited_arxiv_id":null,"evidence_quote":"introduces the rotated Hartree approach that this method resembles but simplifies by deriving the transformation from perturbation theory instead of variational parameters."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the coherent-state formalism and factorization properties used to evaluate Eq. (7)."},{"cited_title":"Sinatra, C","cited_arxiv_id":null,"evidence_quote":"establishes the limited validity time of the truncated Wigner approximation, motivating a beyond-mean-field method for longer times."},{"cited_title":"Pigneur, T","cited_arxiv_id":null,"evidence_quote":"provides the measured relaxation to a phase-locked equilibrium in a one-dimensional bosonic Josephson junction against which the model's damped oscillations are compared."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"describes the configuration-based MCTDHB benchmark whose practical mode count the scrambled mean-field method seeks to bypass."}],"review_version":1}