{"id":"e56dfc60-eff7-474d-9afc-520581dfe18d","arxiv_id":"2507.10272","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A new measure, non-Gaussian entropy, and a beam-splitter protocol estimate bosonic non-Gaussianity using only a constant number of state copies, avoiding full tomography.","lead":"This paper introduces a new measure, called non-Gaussian entropy, for quantifying how far a bosonic quantum state is from a Gaussian state, together with an experimental protocol for measuring it using only four copies of the state and a few beam splitters. A smart generalist might read it because it offers a much cheaper way to characterize non-Gaussian quantum states, which are important resources for continuous-variable quantum computing and sensing.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1/7 proof assumes finite second moments (Eq. A11) but states no such hypothesis; the iff is unproven for infinite-moment pure states.","rationale":"The reader's weakest_assumption identifies exactly the same gap: Theorem 7's proof assumes a finite second moment to justify the expansion in Eq. (A11). My stress-test confirms that this is the most load-bearing point in the central claim, because Theorem 1 inherits the full weight of Theorem 7, and the 'only if' direction collapses if the fixed-point equation has non-Gaussian solutions without regularity. I do not find a stronger objection: the characteristic-function derivation itself is internally consistent under finite second moments, the entropy-additivity argument for pure states is valid, the swap-test/parity relation is standard, and the mixed-state monotonicity proof checks out. The missing assumption is a genuine rigor gap but a patchable one—adding 'with finite second moments' to the theorem statements and proof would resolve it without changing the protocol, which operates on physical finite-energy states. I also note a secondary algebraic slip in the robustness bound of Theorem 9 (the displayed chain in Eq. A17 appears to yield (1−ε)^4 rather than (1−ε)^2 unless an additional inequality is supplied), but that concerns a derived bound, not the central equivalence. Therefore the reader's CONDITIONAL verdict remains appropriate, and my read does not change it.","tokens_in":23505,"tokens_out":28748,"duration_ms":355005,"concrete_test":"Test whether the characteristic-function fixed-point equation g(x)=g(x/2)^3g(−x/2) admits a positive-definite, non-Gaussian solution with infinite second moment that can arise from a pure bosonic state. Concretely, examine g(x)=exp(−x^2(1+ε sin(2π log_2|x|))) for small ε: determine whether it is positive definite and whether its phase-space distribution satisfies the constraints of a pure-state Wigner function. If such a solution exists, Theorem 1 as stated is false and must be amended to finite second moments; if no such solution exists, add the finite-moment hypothesis to Theorems 1 and 7 and confirm that the existing proof goes through unchanged.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim reduces Gaussianity testing to the purity of ψ⊞ψ via Theorem 1, whose proof relies entirely on Theorem 7. In Appendix A, the proof of Theorem 7 defines g(x)=Ξ_ρ(xξ), calls it a classical characteristic function, and then invokes the second-order expansion (A11), g(x)=1−(ξ^TΓξ/4)x^2+o(x^2). That expansion is valid only if the quadrature distribution has finite second moment. No such assumption is stated in Theorem 1 or Theorem 7. There exist legitimate pure bosonic states with finite norm but infinite energy—e.g., wavefunctions with tails ~|q|^{-3/2}—for which g is not differentiable at 0, so the limiting argument in Eq. (A12) is not justified. This is not cosmetic: the functional equation Ξ_ρ(ξ)=Ξ_ρ(ξ/2)^3Ξ_ρ(−ξ/2), which is the engine of the proof, can admit non-Gaussian fixed points when differentiability at 0 is dropped. Consequently the 'only if' direction of Theorem 1 is unproven as stated. For physical finite-energy states the proof probably goes through, so the protocol's operational claim is likely safe, but the theorem statement and proof need an explicit finite-second-moment hypothesis. There is also a minor typo: g(0)=1, not 0 as printed in the proof.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a family of measures called non-Gaussian entropy, NGE_{α,k}(ψ) = S_α(⊞^k ψ), for pure bosonic states, and proposes a tomography-free protocol that uses four copies of an unknown state, three beam splitters, and parity measurements to estimate NGE_{2,1}(ψ). The central theoretical claim (Theorem 1) is that for a pure state ψ, the subadditive inequality S(ψ ⊞ ψ) ≤ 2S(ψ) holds if and only if ψ is Gaussian; equivalently, ψ ⊞ ψ is pure if and only if ψ is Gaussian. This reduces Gaussianity testing to estimating the purity of a beam-splitter output. The paper also extends the framework to mixed states via a Frobenius-norm measure d_F(ρ) and an α-mutual information MING_α(ρ), proves structural properties such as faithfulness and Gaussian invariance, and supplies analytical and numerical examples for Fock states and two-component cat states under loss, dephasing, displacement, and beam-splitter-angle noise.","tokens_in":23778,"tokens_out":14961,"duration_ms":181354,"significance":"If the central theorem is correct, the proposed protocol is a practically valuable, constant-copy method for detecting and quantifying non-Gaussianity without full state tomography, with constant circuit depth and O(1/ε^2) sample complexity. The explicit noise model and the exact analytical formulas for Fock and cat states are useful contributions that go beyond a purely abstract resource-theoretic statement. However, the main theorem currently lacks a stated regularity hypothesis, and the faithfulness of the higher-order measures NGE_{α,k} for k>1 is asserted rather than proved. These issues are load-bearing for the paper's central claims, although they appear fixable for the physically relevant finite-energy setting.","major_comments":[{"comment":"The proof assumes that g(x) = Ξ_ρ(xξ) is a classical characteristic function with finite second moment so that it admits the expansion g(x) = 1 − (ξ^T Γ ξ / 4) x^2 + o(x^2), but Theorem 7 is stated for arbitrary bosonic states and no finite-second-moment or finite-energy hypothesis is included. For a legitimate pure state with infinite second moments, g need not be differentiable at 0 and the limiting argument in Eq. (A12) is not justified. Since Theorem 1, Theorem 2, Theorem 5, Proposition 6, and Proposition 12 all rely on this proof, the manuscript must either add an explicit finite-second-moment (or finite-energy) assumption to the theorem statements or supply a proof that covers infinite-moment states; the precise regularity conditions needed for the central-limit reasoning borrowed from Refs. [39,73] should also be stated.","section":"Appendix A, proof of Theorem 7, Eqs. (A11)-(A12)"},{"comment":"Faithfulness of NGE_{α,k} for general k is justified only by the sentence that '⊞^k ψ is a pure state iff ψ is a Gaussian state,' but no proof of this statement is given. Theorem 1 establishes only the k=1 case, i.e., ψ ⊞ ψ is pure iff ψ is Gaussian. For k>1, the state ⊞^{k-1}ψ is in general mixed when ψ is non-Gaussian, so the claimed equivalence is not a direct corollary of Theorem 1 and requires either an induction argument or a separate proof. This is necessary because these measures are presented as valid non-Gaussianity measures for all k.","section":"Definition 3, Proposition 4(1), Appendix A, Proposition 11(1)"}],"minor_comments":[{"comment":"In the paragraph defining g(x), the text says g(0) = 0; it should read g(0) = 1.","section":"Appendix A, proof of Theorem 7"},{"comment":"The noisy beam-splitter channel integral is written with measure dϕ but the integrand uses the variable φ; the notation should be unified.","section":"Eq. (D1)"},{"comment":"The superscripts containing the Wigner D-matrix arguments have unbalanced parentheses and braces, making these formulas difficult to parse; please rewrite them cleanly.","section":"Appendix E, Eqs. (E9)-(E10)"},{"comment":"The figure references are inconsistent: the text refers to 'Fig. C(a)' and 'Fig. C(c)' where the displayed figure appears to be Fig. 2 with panels (a)-(d); please renumber the cross-references.","section":"Section II.C and Appendix E"},{"comment":"The symbol ε is used both for the target estimation precision in the sample-complexity discussion and for the noise strength via ε_p = e^ε; please use distinct symbols to avoid confusion.","section":"Section II.C"},{"comment":"The phrase 'the maximal overlap with the Gaussian state' should be 'the maximal overlap with a Gaussian state,' since the maximizing Gaussian state is not unique in general.","section":"Theorem 2"}],"recommendation":"major_revision","confidential_remarks":"The central idea is appealing and the protocol is likely correct for finite-energy states, which are the experimentally relevant ones. The main revision should add the missing regularity hypotheses to Theorems 1, 2, 5, and 7 and either prove or cite a proof of the k>1 faithfulness claim. I do not see a reason to reject, but the current theorem statements are stronger than what the supplied proofs establish."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper gives you a genuinely practical tool: a constant-copy, tomography-free way to certify non-Gaussianity of a bosonic state by measuring the purity of a beam-splitter output. The specific move—define non-Gaussian entropy via quantum convolution and read off the answer with parity measurements—is natural but not preempted by the qudit work, and the simplification for zero-mean states is a nice touch. The mixed-state Frobenius measure d_F is a reasonable stopgap, and the authors are upfront that it is not a monotone under Gaussian channels.\n\nWhat the paper does well: the appendices contain real work. There are analytical expressions for Fock-state and cat-state convolution outputs, a clean commutation lemma for Gaussian unitaries and channels, and numerical simulations that include loss, dephasing, and random displacement. None of that is machine-checked, but it is reproducible and gives referees something concrete to verify.\n\nThe soft spots, in order. The proof of Theorem 7, which carries Theorem 1, invokes the second-order expansion of the characteristic function at zero, Eq. (A11). The text says “when the second moment is finite” but never states that as a hypothesis on the state. The stress-test note is right: for infinite-energy pure states the expansion can fail, and the functional equation underlying the proof could in principle admit non-Gaussian fixed points. So the “iff” is unproven as stated. The fix is easy—add finite second moments, or restrict to finite-energy states, which covers every realistic experiment. I do not see this as threatening the protocol's operational claim, but it is a genuine gap in the theorem statement. There is also a minor typo: g(0)=1, not 0.\n\nThe numerics are single-mode and have no error bars, and the noise model has several free parameters. That limits how strongly you can claim experimental readiness, but it does not undercut the idea. The citation pattern is fine; the prior qudit convolution results are cited, and the bosonic extension is distinct.\n\nBottom line: this deserves a serious referee. It is a useful protocol paper with a clean central idea and a proof that needs tightening, not rejection. I would send it out with a request to add the missing regularity hypothesis and to comment on the infinite-moment edge case.","headline":"A clean constant-copy protocol for witnessing bosonic non-Gaussianity, with a theorem that needs one explicit regularity hypothesis before it is fully correct.","tokens_in":24298,"tokens_out":3058,"would_cite":true,"duration_ms":35504,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P45","81P40","81V80"],"pacs":[],"model":"deepseek-v4-flash","headline":"A pure bosonic state is Gaussian exactly when mixing it with itself at a 50:50 beam splitter keeps the entropy subadditivity intact; the paper turns this into a four-copy parity test and a non-Gaussian entropy measure.","keywords":["non-Gaussianity","bosonic states","quantum convolution","beam splitter","parity measurement","swap test","continuous variables","Rényi entropy"],"falsifier":"Search for a pure non-Gaussian state whose characteristic function has a heavy tail or unbounded second moment and compute whether $\\mathrm{Tr}[(\\psi \\boxplus \\psi)^2] < 1$; numerically, minimize the purity of the beam-splitter output over a family of such states to see whether any candidate drives the purity to 1. Finding a single pure non-Gaussian $\\psi$ with $\\mathrm{Tr}[(\\psi \\boxplus \\psi)^2] = 1$ would refute Theorem 1; finding none would confirm it, and the search would also reveal whether the finite-second-moment assumption is the true boundary of the theorem.","tokens_in":23292,"feed_emoji":"⚛️","tokens_out":8322,"duration_ms":89114,"temperature":0.7,"pith_summary":"This paper aims to make the question \"is a bosonic state non-Gaussian?\" answerable by a measurement that uses only a constant number of copies. Its central claim is that for any pure bosonic state $\\psi$, the subadditivity inequality $S(\\psi \\boxplus \\psi) \\le 2S(\\psi)$ holds if and only if $\\psi$ is Gaussian, where $\\boxplus$ is the quantum convolution produced by sending two copies through a 50:50 beam splitter. Because $\\psi \\boxplus \\psi$ is pure exactly when $\\psi$ is Gaussian, detecting non-Gaussianity reduces to estimating the purity of the beam-splitter output, and a parity measurement on that output provides the estimate directly. The paper introduces the non-Gaussian entropy $NGE_{\\alpha,k}(\\psi) = S_\\alpha(\\boxplus^k \\psi)$ as a resource measure and gives a four-copy, three-beam-splitter protocol to measure it, with a three-copy simplification for zero-mean states. If correct, this gives an experimentally accessible, tomography-free way to quantify non-Gaussianity in continuous-variable systems.","feed_headline":"One parity count on four copies flags non-Gaussian states","feed_subtitle":"If two copies stay pure through a beam splitter, the state is Gaussian; any impurity is a measurable non-Gaussian signature.","key_machinery":"The load-bearing object is quantum convolution, $\\rho \\boxplus \\sigma := \\mathrm{Tr}_B[U_{\\pi/4}(\\rho \\otimes \\sigma)U_{\\pi/4}^\\dagger]$, the reduced state of one output port of a 50:50 beam splitter. Two identities carry the argument. First, the beam splitter maps the symmetric and antisymmetric components of two identical inputs to even- and odd-parity subspaces, giving $\\mathrm{Tr}[\\rho\\sigma] = \\mathrm{Tr}[(\\rho \\boxminus \\sigma)P]$, so the purity of $\\psi \\boxplus \\psi$ is read off from a parity expectation. Second, iterated convolution factorizes the characteristic function as $\\Xi_{\\boxplus^k \\psi}(\\xi) = \\Xi_\\psi(\\xi/\\sqrt{m_k})^{a_k} \\Xi_\\psi(-\\xi/\\sqrt{m_k})^{b_k}$ with $m_k = 4^k$, and a quantum central-limit argument shows the iterates converge to a Gaussian; this is what forces the entropy inequality to be tight only for Gaussian states.","core_discovery":"The core discovery is an equivalence between Gaussianity and the saturation of a classical-looking entropy inequality under quantum convolution. For a pure state $\\psi$, the paper proves $S(\\psi \\boxplus \\psi) \\le 2S(\\psi)$ is tight if and only if $\\psi$ is Gaussian; every non-Gaussian pure state violates it. Equivalently, the output of a 50:50 beam splitter fed with two copies of $\\psi$ is a pure state if and only if $\\psi$ is Gaussian. This converts Gaussianity testing into a purity test: the average parity $\\langle P \\rangle = \\mathrm{Tr}[(\\psi \\boxplus \\psi)^2]$ equals 1 iff $\\psi$ is Gaussian, and if the maximal fidelity to a Gaussian state is $1-\\epsilon$, then $\\langle P \\rangle \\ge (1-\\epsilon)^2$. The same idea extends to mixed states: $\\rho$ is Gaussian iff $U_{\\pi/4}(\\rho \\otimes \\rho)U_{\\pi/4}^\\dagger$ is a product state, which motivates the measurable Frobenius measure $d_F(\\rho) = \\lVert \\rho_{AB} - \\rho_A \\otimes \\rho_B \\rVert_F$ and the mutual-information-based non-Gaussianity measure.","pith_inferences":["A promising extension is to use the same parity primitive as a certification tool in boson-sampling and bosonic error-correction experiments, where tomography is infeasible but a constant-copy purity check is realistic.","The proof's finite-second-moment assumption suggests the sharp Gaussian threshold may fail for states with heavy-tailed characteristic functions; a targeted search over such states would map the exact domain of the theorem.","If the non-Gaussian entropy behaves like other resource measures, it could provide the missing link between environmental non-Gaussianity and the quantum capacity of beam-splitter channels, a question the paper leaves open.","Because $d_F$ is not contractive under Gaussian operations, applying it as a resource quantifier requires care; a natural follow-up is to pair the efficient parity estimator with a contractive relative-entropy measure estimated on the relevant state family."],"forward_implications":["An unknown pure bosonic state can be certified and quantified as non-Gaussian with four copies (three for zero-mean states) using only beam splitters and parity measurements, sidestepping full state tomography.","Estimating the parity to additive error $\\varepsilon$ requires $O(1/\\varepsilon^2)$ copies and, for $N$-mode inputs, $O(N)$ beam-splitter gates at constant circuit depth, so the test remains practical for many-body systems.","The Rényi non-Gaussian entropies satisfy faithfulness, Gaussian invariance, and additivity under tensor products, making them legitimate measures of non-Gaussianity for pure states.","For mixed states, the Frobenius measure $d_F$ is faithful and measurable by swap-test circuits, though the paper explicitly notes it is not a monotone under Gaussian operations.","Numerical simulations with cat states and Fock states under loss, dephasing, random displacement, and noisy beam splitters show the parity signal persists at weak noise."],"supporting_citations":[{"why":"Supplies the quantum central limit theorem used to show iterated convolution converges to a Gaussian.","marker":"[39]"},{"why":"Provides the Gaussianity identification used after the limiting characteristic function is shown to be Gaussian.","marker":"[73]"},{"why":"Earlier quantum entropy and central limit work that frames convolution-based entropy measures.","marker":"[42]"},{"why":"Demonstrates beam-splitter interference with parity measurement to reveal the purity or overlap of bosonic states.","marker":"[48]"},{"why":"Direct estimation of nonlinear functionals of quantum states, underlying the swap-test purity estimate.","marker":"[50]"}],"fun_headline_variants":["Gaussian if two copies stay pure through a beam splitter","Four copies and three splitters measure non-Gaussian entropy","Non-Gaussian states fail a simple purity test","Beam-split purity test flags non-Gaussian states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof that the inequality is tight exactly for Gaussian states assumes the state's characteristic function is that of a classical probability law with finite second moment, so the Taylor expansion used in the limiting argument is valid, and it assumes the quantum central-limit convergence step applies to the state under test.","fun_headline_variants_meta":{"raw":{"variants":["Gaussian if two copies stay pure through a beam splitter","Four copies and three splitters measure non-Gaussian entropy","Non-Gaussian states fail a simple purity test","Beam-split purity test flags non-Gaussian states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000348,"raw_usage":{"total_tokens":1896,"prompt_tokens":930,"completion_tokens":966,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":900}},"tokens_in":546,"tokens_out":966,"duration_ms":8900,"temperature":1.0,"reasoning_tokens":900,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T17:33:39.853279+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search for a pure non-Gaussian state whose characteristic function has a heavy tail or unbounded second moment and compute whether $\\mathrm{Tr}[(\\psi \\boxplus \\psi)^2] < 1$; numerically, minimize the purity of the beam-splitter output over a family of such states to see whether any candidate drives the purity to 1. Finding a single pure non-Gaussian $\\psi$ with $\\mathrm{Tr}[(\\psi \\boxplus \\psi)^2] = 1$ would refute Theorem 1; finding none would confirm it, and the search would also reveal whether the finite-second-moment assumption is the true boundary of the theorem.","supporting_citations":[{"cited_title":"Soto and P","cited_arxiv_id":null,"evidence_quote":"Supplies the quantum central limit theorem used to show iterated convolution converges to a Gaussian."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Gaussianity identification used after the limiting characteristic function is shown to be Gaussian."},{"cited_title":"Moura Alves and D","cited_arxiv_id":null,"evidence_quote":"Direct estimation of nonlinear functionals of quantum states, underlying the swap-test purity estimate."}],"review_version":1}