{"id":"54438d4f-4f47-46e9-a0c7-1fc1daef043c","arxiv_id":"2508.05352","paper_version":1,"verdict":"REJECT","confidence":"LOW","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":6,"one_line_summary":"A linear-depth, 30-qubit geometric encoding generates a Re = 35,000 turbulent field on 1024^3 points, but its headline k^{-5/3} spectrum is built into the encoding and the submitted abstract belongs to a different paper.","lead":"Researchers propose generating turbulent fluid fields directly inside quantum circuits, replacing the costly step of copying billions of grid points with a geometric encoding that runs on 30 qubits at Reynolds number 35,000. A caveat: the Kolmogorov spectrum that validates the method is baked into the encoding, and the submitted abstract describes a different, recommender-system paper.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (9)'s linear-angle ansatz has no fidelity bound, and the k^{-5/3} demonstration is self-referential; the Θ(n)/log-Re claim is therefore unsecured.","rationale":"The submitted artifact is internally mismatched: the title, abstract, and category describe an M3BSR recommendation paper, while the full text is the turbuloscope quantum-encoding paper. Following the instruction to treat all parts as in-scope, I weight this mismatch heavily for the submitted artifact itself, but the strongest_claim the reader extracts is the full-text state-preparation claim. Stress-testing that claim, the most load-bearing step is Eq. (9): it is the single mechanism that reduces parameter count and circuit depth. Without a fidelity guarantee, the Θ(n) and log-Re assertions do not follow from the text. The demonstration is self-referential because the amplitude target already encodes k^{-5/3}, and the convolution is said not to change the spectrum, so the reported spectrum cannot independently confirm the hyperplane approximation. The missing SI, single-realization statistics, and NISQ infeasibility compound the problem but are less fundamental. This matches the reader's weakest_assumption. Since the gap is serious enough to leave the central claim unsupported, keeping REJECT is appropriate; hence UNCHANGED.","tokens_in":29692,"tokens_out":5400,"duration_ms":62381,"concrete_test":"Using the released QEncodeTurb code, compute the exact conditional rotation angles θ_j for a Gray-coded A(k)∝k^{-5/3} target at n=15, 21, and 30 (or as large as feasible), and compare them with the Eq. (9) ridge-regression predictions from the paper. Report the L2 and max angle errors, the resulting squared fidelity between the approximate and target amplitude states, and the wavenumber-binned amplitude error. If fidelity falls below ~0.99 at n=30 or the error grows with n, the linear ansatz is not faithfully preparing the target distribution and the log-Re scaling claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The full text's central claim is that turbulence state preparation is Θ(n)-depth and log-Re with no ancillas. The entire compression from O(2^n) parameters to O(n^2) rests on Eq. (9): θ_j(q_<j) ≈ b_j + Σ w_{j,m} q_m, justified only by an asserted 'smooth low-curvature manifold' in Gray-code feature space. No theorem or numerical fidelity bound is supplied for this approximation; Methods say only that amplitude-weighted ridge regression gives a closed-form solution. Without a bound relating regression loss to final-state fidelity, there is no reason a 10n-depth circuit prepares the target A(k)∝k^{-γ}; the approximation error could concentrate in the exponentially many high-wavenumber basis states. The reported k^{-5/3} spectrum cannot validate this, because that spectrum is the stage-1 input target (Fig. 1e), and Fig. 2b states it is nearly unchanged by convolution. Intermittency metrics are also computed on a single generated field with no error bars or DNS baseline. The deferred SI [46] is said to contain the optimality proof and measurement oracles, so the core derivations are not checkable in the submitted text. Discussion concedes direct NISQ execution is prohibitive; however the key unsecured step remains Eq. (9). If this approximation degrades with n or fidelity, the Θ(n)-depth and log-Re scaling claims collapse.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The submitted front matter is internally inconsistent: the title and abstract describe a multi-modal multi-behavior sequential recommendation model (M^3BSR), whereas the body is a quantum-computing paper, “Geometric encoding of turbulence for end-to-end quantum simulation,” proposing a three-stage “turbuloscope” state-preparation protocol. I evaluate the body as the substantive scientific content. The paper claims a Θ(n)-depth quantum circuit with no ancillary qubits that encodes a k^{-5/3} turbulent field on a 1024^3 grid with 30 qubits (Re ≈ 35,000), with circuit depth scaling linearly in n and Reynolds number scaling logarithmically. The Methods combine Gray-code basis ordering, a linear ansatz for conditional rotation angles, random phase scrambling, a spectral convolution measurement, and a classical deconvolution step. The demonstration reports Kolmogorov k^{-5/3} energy scaling, a k^{-11/3} vortex-surface-field spectrum, a stretched-exponential vorticity PDF, and structure-function exponents close to SL94.","tokens_in":30030,"tokens_out":5610,"duration_ms":68853,"significance":"If the central claims held, the work would remove a recognized bottleneck in quantum fluid simulation: preparing a high-Reynolds-number turbulent initial state with shallow circuits and logarithmic qubit scaling. The manuscript includes useful elements: a closed-form, non-iterative parameter-fitting procedure; public code and data links; and a transpiled-depth scaling plot (Fig. 4c). However, the validity of the main claim is not currently established. The key compression assumption, Eq. (9), has no fidelity bound; the headline spectral law is an input of the construction rather than an emergent validation; the ancilla-free claim conflicts with the measurement protocol; and the statistical evidence rests on a single field with no error bars or DNS baseline. The significance is therefore potential rather than demonstrated.","major_comments":[{"comment":"The Θ(n)-depth and O(n^2)-parameter claims collapse if the linear ansatz in Eq. (9) is not a sufficiently accurate approximation, yet no error bound is given relating the amplitude-weighted ridge-regression residual to the fidelity of the prepared state against the exact target A(k) ∝ k^{-γ}. The text states only that the manifold is “smooth” and that regression is amplitude-weighted; no analysis is provided for n → ∞, for different γ, or for the high-wavenumber modes that dominate the inertial range. Without such a bound, the exponential-to-polynomial parameter compression is unsupported.","section":"Methods, Eq. (9)"},{"comment":"The k^{-5/3} energy spectrum in Fig. 3c is not an emergent validation: the amplitude-encoding stage is explicitly designed to set a specified power-law spectrum (Results: “facilitates setting a specified power-law spectrum”; Fig. 1e fixes γ = 5/3), and Fig. 2b states that the convolution leaves the angle-averaged spectrum nearly unaffected. The reported E(k) ≈ k^{-5/3} is therefore a check on the input spectrum, not evidence that the protocol generates turbulent physics. The VSF k^{-11/3} scaling and the PDF/structure-function diagnostics are computed on the same generated field and inherit this limitation.","section":"Fig. 1e, Fig. 2b, Fig. 3c"},{"comment":"The headline claim that the algorithm “requires no ancillary qubits” is contradicted by the spectral measurement protocol, which uses an ancilla-assisted Hadamard test (Methods: “Measurement of spectral observables”) and a post-selection step in the complexity analysis (Results: “We analyze the computational complexity... a post-selection step with a success probability converging to 1”). If the no-ancilla statement is intended only for state preparation, that scope must be stated. Moreover, the convergence of the post-selection success probability to 1 is asserted without proof, and it directly affects the claimed end-to-end complexity.","section":"Abstract vs. Methods: Measurement of spectral observables"},{"comment":"The statistical validation is based on a single generated field. No error bars are provided for E(k), for the stretched-exponential PDF parameters (σ = 0.05, β = 0.5), or for the structure-function exponents; there is no comparison to a direct numerical simulation or to a classical synthetic field with the same energy spectrum. Consequently, the claims of “strong intermittency,” agreement with SL94, and statistical isotropy (anisotropy tensor of order 10^{-3}) are not quantitatively supported.","section":"Results: Quantum encoded instantaneous turbulent field, Fig. 3"},{"comment":"Core derivations are deferred to the Supplemental Information: the definitions of C, m(κ,κ′), n(κ′), the summation set in Eq. (5), the measurement oracles, the theoretical lower bound, and the proof of asymptotic optimality. As submitted, the main text does not contain a verifiable proof of the central Θ(n)-depth or log-Re scaling claims. The reader is asked to accept the central claim on the basis of deferred material and a small-scale classical simulation.","section":"SI [46] and Methods, Eq. (5)"}],"minor_comments":[{"comment":"There is a typo in the text: “assess the the statistics” should be “assess the statistics.”","section":"Methods: Measurement of mean momentum"},{"comment":"The distinction between “unoptimized CZ-gate count” and “transpiled circuit depth” needs clarification; provide the concrete gate count for n = 30 and the assumed connectivity, error model, and optimization settings used for transpilation.","section":"Fig. 4c"},{"comment":"The GitHub links are welcome, but no version, license, or reproducibility instructions are given; a specific commit hash and environment description would help.","section":"Data and code availability"},{"comment":"The title and abstract of the submission do not match the body text. This is not a purely cosmetic issue; it makes the manuscript incoherent as a submission and must be resolved editorially.","section":"Front matter"}],"recommendation":"reject","confidential_remarks":"The submitted package is seriously incoherent: the title/abstract describe a recommendation-system paper while the body is a quantum-turbulence paper. Even setting that aside, the central scientific claim (Θ(n)-depth, ancilla-free, log-Re state preparation) is unsupported: Eq. (9) has no fidelity bound, the spectral validation is largely circular, the ancilla-free claim conflicts with the measurement protocol, and the statistical evidence lacks baselines and error bars. These are load-bearing defects, not local presentation issues. I do not see how they can be fixed within the scope of the current manuscript; a full rewrite with new validation would be needed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Let me give you the short version: this submission is not reviewable as it stands. The abstract, title, authors, and category all describe a multi-modal multi-behavior recommendation model (M^3BSR), but the full text is a quantum encoding paper by different authors. The abstract's claim that M^3BSR outperforms state-of-the-art recommenders has zero supporting text. That mismatch alone is a desk reject, and it is decisive.\n\nEven so, I read the actual body carefully, because that is where the real content lives. The genuine novelty is the three-stage protocol: Gray-code amplitude encoding using a linear ansatz, random phase scrambling, and a Hopf-fibration-based mapping of observables to vortex tubes, all aimed at preparing turbulent states in Θ(n) circuit depth with log-Re qubits. The Gray-code locality idea is a sensible way to smooth a power-law amplitude, and the geometric story is coherent. The paper also ships code and data on GitHub, which is real evidence of reproducibility for the demonstrations. So there is actual work here.\n\nThe soft spots are severe, though. The headline validation is circular: the energy spectrum E(k) ~ k^{-5/3} is set directly as the amplitude target in stage 1 (γ = 5/3 in Fig. 1e), and the convolution is stated to leave the angle-averaged spectrum nearly unaffected (Fig. 2b). Reporting that as \"reproducing Kolmogorov's 5/3 scaling\" is taking credit for an input. The intermittency statistics come from the same single generated field, with no error bars and no DNS comparison. More fundamentally, Eq. (9) — the linear ansatz for rotation angles — is the entire compression from exponential to O(n^2) parameters, yet no fidelity bound is given. The paper asserts a smooth manifold in Gray-code space; if that approximation degrades with n, the Θ(n)-depth claim collapses. And the asymptotic-optimality proof, the ridge-regression derivation, and the measurement oracles are all deferred to an SI that is not in the submission. I cannot verify the main claim. To the authors' credit, the Discussion openly concedes that direct NISQ execution is prohibitive, but that does not fill the missing proof.\n\nThe right move is a desk reject with feedback: fix the front matter, supply the missing SI and error bounds, and either avoid the circularity or reframe the spectrum result as a target check rather than a physics reproduction. This version is not ready for referees, but the underlying idea is worth watching. I would not cite it yet, though I would bring it up in a reading group if someone wants a good discussion about what counts as validation in quantum state preparation.","headline":"The submission is a broken artifact: the front matter advertises a recommendation paper, the body is a quantum turbulence paper, and the body's central scalability claim rests on an unproven linear ansatz plus a circular spectrum reproduction—but the underlying idea is worth engaging with if properly resubmitted.","tokens_in":30612,"tokens_out":5211,"would_cite":false,"duration_ms":59692,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A quantum encoding method called turbuloscope can prepare a fully turbulent flow with only 30 qubits, using a linear-depth circuit and no auxiliary qubits.","keywords":["quantum state preparation","turbulence","multiscale encoding","Gray code","Hopf fibration","amplitude encoding","Reynolds number scaling","intermittency"],"falsifier":"Encode a pure power-law spectrum with $n=51$ qubits using the linear-ansatz circuit, reconstruct the amplitudes of all basis states, and compare against the exact target $A(k)\\propto k^{-5/3}$; if the weighted reconstruction error grows faster than polynomially in $n$, or if the recovered energy spectrum deviates from $k^{-5/3}$ across the inertial range, the central scaling claim is falsified.","tokens_in":29426,"feed_emoji":"🌀","tokens_out":15244,"duration_ms":143915,"temperature":0.7,"pith_summary":"This paper claims that quantum state preparation for turbulence does not have to be the exponential I/O bottleneck it is usually taken to be. It introduces turbuloscope, a three-stage geometric encoder that builds an instantaneous turbulent field directly from the self-similar, power-law structure of the energy cascade: a Gray-code basis makes conditional rotation angles smooth, a linear hyperplane fit sets them from $O(n^2)$ parameters, and a Hopf-fibration mapping turns quantum observables into vortex tubes. The resulting no-ancilla circuit has depth $\\Theta(n)$, and the qubit count grows logarithmically with Reynolds number; the demonstration encodes a $1024^3$ field at $\\mathrm{Re}\\approx 35{,}000$ with 30 qubits, reproducing the $k^{-5/3}$ spectrum and intermittency statistics. If correct, this removes the main obstacle to using quantum simulation for high-Reynolds-number engineering flows and other multiscale systems.","feed_headline":"Thirty qubits encode a billion-point turbulent flow","feed_subtitle":"The new encoding grows qubit count only logarithmically with Reynolds number, clearing the state-preparation bottleneck.","key_machinery":"The load-bearing machinery is the 'turbuloscope' encoding pipeline. Its algebraic core is the linear ansatz $\\theta_j(\\mathbf{q}) \\approx b_j + \\sum_{m<j} w_{jm} q_m$, which makes the exponential state-preparation problem polynomial by flattening the rotation-angle landscape in a Gray-code basis; the parameters are fixed by a closed-form amplitude-weighted ridge regression. The geometric core is the generalized Madelung transform together with the Hopf fibration, which sends the unit spin vector of the prepared quantum state to vorticity in physical space, with each point of the Bloch sphere corresponding to a vortex line and each patch to a vortex tube. These pieces combine into a three-sta","core_discovery":"The paper's central claim is that the data-loading bottleneck for turbulent flows can be bypassed by exploiting scale invariance. For a target spectrum $E(k)\\propto k^{-\\gamma}$, the conditional rotation angles $\\theta_j$ that build the amplitude distribution are approximated by a linear ansatz $\\theta_j(\\mathbf{q})\\approx b_j+\\sum_{m<j}w_{jm}q_m$ in Gray-code feature space, with parameters obtained in closed form by amplitude-weighted ridge regression. This compresses the exponential number of controlled rotations to $O(n^2)$ parameters and yields a linear-depth circuit that prepares a complex-valued turbulent state without ancillas. A phase-scrambling layer adds spatial correlations, and o","pith_inferences":["Document note: the metadata and abstract supplied with this text describe a different manuscript; the claims above follow the full text, which is the quantum geometric-encoding paper.","Editorial inference: the same Gray-code-plus-hyperplane recipe is also a classical compression algorithm—it produces a full turbulent field from $O(n^2)$ parameters—so the approach has value as a generative model even without quantum hardware.","Editorial inference: the fidelity of the linear ansatz is demonstrated for one spectrum exponent at one resolution; sweeping the exponent and the number of qubits while tracking reconstruction error would reveal how far the $\\Theta(n)$ scaling extends.","Editorial inference: the practical near-term path may be hybrid—prepare or verify the initial field classically, then let a quantum processor evolve it—since direct execution today is constrained by all-to-all connectivity and coherence limits."],"forward_implications":["Turbulent initial conditions for quantum PDE solvers can be prepared in $\\Theta(n)$ depth, replacing the $\\Theta(2^n/n)$ cost of optimal general data-loading.","Reynolds number scales as $\\mathrm{Re}\\sim 2^{4n/9}$: each additional qubit multiplies the accessible Reynolds number by a constant factor, so going from 30 to 51 qubits moves the reachable regime from $\\mathrm{Re}\\approx 35{,}000$ to beyond $10^7$.","The prepared state has coherent vortex tubes and reproduces both energy and vortex-surface spectral laws, so it can serve as a physically faithful initial condition for Hamiltonian simulation rather than a spectral fake.","The circuit parameters come from a one-shot closed-form regression, so the state-preparation protocol avoids iterative variational optimization and its convergence issues.","Because the primitive only exploits self-similarity and locality, the same method transfers to other power-law multiscale systems, including magnetohydrodynamic turbulence, cosmic structure, and reaction-diffusion patterns."],"supporting_citations":[{"why":"Establishes the optimal circuit-depth lower bound for general quantum state preparation that the linear-depth encoding claims to bypass.","marker":"[18]"},{"why":"Provides the asymptotically optimal depth scaling for general unitary synthesis, used as the comparison baseline for the claimed speedup.","marker":"[19]"},{"why":"Supplies the Hopf fibration construction that links the geometry of linked and knotted fields to vortex-tube structures.","marker":"[26]"},{"why":"Defines the generalized Madelung transform connecting wavefunction dynamics to fluid density and momentum, the basis for the observable convolutions.","marker":"[27]"},{"why":"Introduces the quantum spin representation that supplies the spin-vector-to-vorticity mapping used in the encoding.","marker":"[28]"},{"why":"Provides the linear circuit ansatz that the amplitude-encoding stage is rooted in.","marker":"[34]"},{"why":"Documents Gray-code versus binary locality, the reason the conditional rotation angles become smooth enough for the linear fit.","marker":"[39]"},{"why":"The SL94 model whose scaling exponents are used to validate the generated field's intermittency.","marker":"[41]"},{"why":"Reference DNS study of high-Reynolds-number isotropic turbulence providing the coherent-structure and statistical benchmarks for comparison.","marker":"[49]"}],"fun_headline_variants":["Diffusion-based denoising enhances multi-modal sequential recommendations","M^3BSR: Conditional diffusion cleans noise in sequential recommendations","Denoising user behavior and item modalities boosts recommendation quality","Conditional diffusion removes noise for better multi-modal recommendations","New model denoises behavior and modality to sharpen sequential recommendations"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"All the claimed speedup rests on the assumption that the conditional rotation angles needed to build a power-law amplitude spectrum are a smooth, nearly linear function of the previously encoded bits in Gray-code order, so that a single closed-form linear regression captures the whole field; if that approximation degrades with grid size or fidelity, the claimed $\\Theta(n)$ depth and logarithmic-in-Reynolds-number scaling collapse.","fun_headline_variants_meta":{"raw":{"variants":["Diffusion-based denoising enhances multi-modal sequential recommendations","M^3BSR: Conditional diffusion cleans noise in sequential recommendations","Denoising user behavior and item modalities boosts recommendation quality","Conditional diffusion removes noise for better multi-modal recommendations","New model denoises behavior and modality to sharpen sequential recommendations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000768,"raw_usage":{"total_tokens":3260,"prompt_tokens":780,"completion_tokens":2480,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":524,"completion_tokens_details":{"reasoning_tokens":2398}},"tokens_in":524,"tokens_out":2480,"duration_ms":22357,"temperature":1.0,"reasoning_tokens":2398,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T23:25:33.267685+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Encode a pure power-law spectrum with $n=51$ qubits using the linear-ansatz circuit, reconstruct the amplitudes of all basis states, and compare against the exact target $A(k)\\propto k^{-5/3}$; if the weighted reconstruction error grows faster than polynomially in $n$, or if the recovered energy spectrum deviates from $k^{-5/3}$ across the inertial range, the central scaling claim is falsified.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the asymptotically optimal depth scaling for general unitary synthesis, used as the comparison baseline for the claimed speedup."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Hopf fibration construction that links the geometry of linked and knotted fields to vortex-tube structures."},{"cited_title":"Meng and Y","cited_arxiv_id":null,"evidence_quote":"Defines the generalized Madelung transform connecting wavefunction dynamics to fluid density and momentum, the basis for the observable convolutions."},{"cited_title":"Meng and Y","cited_arxiv_id":null,"evidence_quote":"Introduces the quantum spin representation that supplies the spin-vector-to-vorticity mapping used in the encoding."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents Gray-code versus binary locality, the reason the conditional rotation angles become smooth enough for the linear fit."},{"cited_title":"She and E","cited_arxiv_id":null,"evidence_quote":"The SL94 model whose scaling exponents are used to validate the generated field's intermittency."},{"cited_title":"Ishihara, T","cited_arxiv_id":null,"evidence_quote":"Reference DNS study of high-Reynolds-number isotropic turbulence providing the coherent-structure and statistical benchmarks for comparison."}],"review_version":1}