{"id":"b27172d1-0508-48c1-92e0-9d6129ad84f3","arxiv_id":"2512.19457","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The optimal Gaussian entropic transport plan is obtained by shrinking the spectrum of the aligned correlation operator with f_ε(x) = 2x/(√(4x²+ε²)+ε), and its ε→0 limit is the most diffusive (centroid) optimal plan.","lead":"Entropic optimal transport between Gaussian measures is shown to be a spectral shrinkage: the optimal coupling applies one universal scalar function to the eigenvalues of the aligned correlation operator, covering degenerate and infinite-dimensional settings. This yields exact ε→0 selection, convergence rates, and a direct algebraic evaluation of entropic costs across all regularization levels.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The paper's 'no assumptions' universality and the Monge-rejection selection principle rest entirely on degenerate-case machinery imported from the author's own unreviewed preprint [40]; if [40, Thm 2.1] or the Schur-complement criterion is defective, Theorems 3.4 and 3.9 collapse.","rationale":"The reader's weakest assumption identifies exactly the load-bearing dependence on [40]'s degenerate-case results. Without a properly aligned Green pair, Theorem 3.4's optimization argument cannot even be stated; without the full N12-parameterization, the claimed selection principle and 'never Monge' outcome are unjustified. The present paper's own proof of Theorem 3.4 onwards is internally coherent given that input: the spectral argument is sound, Lemma A.1 is valid, and the cost/bias computations follow from the closed form. Thus the central mathematical claim is credible but only conditionally. The algorithmic overstatement (O(d) per ε for the full coupling) is a secondary issue, not the load-bearing one. Re-deriving [40]'s Theorem 2.1 and the Schur-complement criterion would settle the concern; until then the correct verdict remains CONDITIONAL. My read does not move the reader's verdict.","tokens_in":24242,"tokens_out":10523,"duration_ms":107121,"concrete_test":"Independently re-derive [40, Theorem 2.1] and the B/A=0 uniqueness criterion from first principles, then check the family of degenerate pairs A=diag(I_r,0), B block B11, B12, B22 with r=1,2 and infinite-dimensional H2. Specifically: (1) verify every M of the stated form is a properly aligned Green operator; (2) verify every optimal coupling covariance (2) arises from some such M (completeness); (3) test reachability when dim(R(B1/2)∩H2) > dim(N(B1/2A1/2)∩H1), confirming A⇝B fails and B⇝A holds; (4) confirm B/A=0 iff the optimal coupling is unique. If any step fails, the universal claim of Theorem 3.4 and the 'never Monge' conclusion of Theorem 3.9 need qualification.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 3.4 promises a unique Gaussian EOT coupling for arbitrary trace-class A,B, with no nullity or inclusion assumptions. The proof requires a properly aligned Green pair (G,M); the existence of such a pair for arbitrary A,B is not proved here but is taken from [40]. More heavily, Theorem 3.9's selection principle ('never Monge' when B/A≠0, convergence to the N12=0 centroid) depends on [40, Theorem 2.1], which claims to characterize ALL optimal Kantorovich couplings via the free parameter N12 subject to N12*N12 ⪯ B/A, and on the stated equivalence B/A=0 ⇔ uniqueness. None of these structural theorems is reproduced in this paper. If [40, Thm 2.1] is incomplete—e.g., if there exist optimal couplings not representable in that form, or if proper alignment fails in some infinite-dimensional degenerate case—then the universal spectral-shrinkage formula and the central claim that the entropic limit systematically rejects Monge solutions are unsupported. This is not a mere citation gap: the paper's abstract and discussion present these as established facts, and the examples (OP, IBM) rely on them. The rest of the derivation—the von Neumann/variational argument, the scalar optimization, and Lemma A.1—is only valid conditional on that imported existence and parameterization.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies entropic optimal transport (EOT) between centered Gaussian measures on a separable Hilbert space with trace-class covariances A and B. It claims that for any properly aligned Green's operators G,M (GG*=A, MM*=B, G*M≥0), the unique optimal Gaussian entropic coupling has cross-covariance Cε=G fε(G*M) M*, with universal spectral shrinkage fε(x)=2x/(√(4x²+ε²)+ε). This is proved by diagonalizing G*M, applying the von Neumann trace inequality, and solving a scalar problem. The paper then derives the entropic cost bias (Cor. 3.5), recovering the finite-rank rate ε log(1/ε) and a non-parametric rate ε^{1−1/α} under exact polynomial spectral decay; examines the ε→0 limit (Thm 3.9), claiming convergence to the most diffusive optimal Kantorovich plan (the N12=0 centroid), which is never a Monge plan when the Schur complement is nonzero; and gives stability bounds and closed-form coupling distances (Thms 3.10, 3.12). Structural facts about degenerate Gaussian OT—existence of properly aligned pairs and parameterization of all optimal couplings—are imported from the author's companion preprint [40].","tokens_in":24420,"tokens_out":14239,"duration_ms":128535,"significance":"If the imported structural results are sound, the spectral-shrinkage formula is a significant unifying result: it removes injectivity and range-inclusion assumptions, gives a transparent geometric interpretation, and reduces computation to one spectral decomposition for all ε. The derivation of Theorem 3.4 is clean and self-contained conditional on the existence of a properly aligned pair; the bias and coupling-distance formulas are explicit, and the finite-rank ε log(1/ε) recovery is reassuring. The predicted non-parametric rates for infinite-dimensional settings are a useful caution. However, the paper's universality and selection principles are only as strong as the unproved results in [40]; the rate theorem's exact power-law assumption also requires scrutiny. With those points addressed, the paper would be a valuable contribution.","major_comments":[{"comment":"The proof of Theorem 3.4 assumes existence of a properly aligned Green pair (G,M) for arbitrary trace-class A,B; this existence is not proved here but imported from [40]. Similarly, Theorem 3.9's identification of the entropic limit with the N12=0 plan and the 'never Monge' conclusion depends entirely on [40, Theorem 2.1], which parameterizes all optimal Kantorovich couplings. Since these structural results are load-bearing for the paper's central universality and selection claims, the manuscript is not self-contained. Please either prove these results in an appendix or explicitly state the theorems under the assumption that such a properly aligned pair exists / that Theorem 2.1 holds, and flag the degree of reliance.","section":"§2.2, Thm 3.4, Thm 3.9"},{"comment":"The nonparametric rate theorem assumes an exact power-law spectrum s_k=c k^{-α}. The Riemann-sum calculation in Cor 3.5(2) uses this exact identity to substitute u_k=k(ε/c)^{1/α}. For integrated Brownian motions, only the asymptotic equivalence s_k≍(πk)^{-n} is available; the exact equality fails at every finite k. Therefore the claimed 'exact asymptotic convergence rate' for IBMs does not follow from the stated proof. A comparison or Tauberian argument is needed to justify the same exponent (and constant) under asymptotic polynomial decay, or the theorem should be restricted to operators with exact power-law spectra.","section":"Cor 3.5(2), Ex 3.1"},{"comment":"The purported constructive update M0=K0+(B−K0K0*)^{1/2} (and the general Remark 3.2) is not proved: it is not shown that the cross term K0(B−K0K0*)^{1/2} vanishes for the coordinate-free K0, nor that the operator square root can be chosen to make M0 a Green's operator for B; the block verification assumes the canonical block decomposition and cites [40, Lemma 2.4]. Since this construction is offered as the route to proper alignment for arbitrary A,B, it needs a full proof or should be removed from the claims of full generality.","section":"Prop 3.8, Remark 3.2"}],"minor_comments":[{"comment":"Typo: 'our result recovers the our result recovers' should read 'our result recovers'.","section":"After Cor 3.5"},{"comment":"In the display for Qε, '(M∗M)' should be '(M0∗M0)'.","section":"Thm 3.10"},{"comment":"Reference [40] is an unpublished preprint; given the heavy reliance, the reference should be marked as 'submitted' or 'preprint' and, ideally, the relevant statements reproduced.","section":"References"},{"comment":"The spectral convention in §2.1 orders degenerate nonzero eigenvalues by magnitude starting at λ2, while Theorem 3.4 writes λ1 ≥ λ2 ≥ ... > 0 with zero excluded; please harmonize the notation.","section":"§2.1/Thm 3.4"}],"recommendation":"major_revision","confidential_remarks":"The manuscript depends on the author's companion preprint [40] for the central structural results; I recommend that the editor require a published or appended version of those proofs before acceptance. The exact-power-law assumption in the rate theorems should also be addressed by the authors."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Victor—short version: this paper is worth refereeing, not one to trust at face value. The core spectral-shrinkage result is a clean repackaging of known Gaussian EOT closed forms, and for full-rank marginals it is correct. What is genuinely new is the coordinate-free setup covering degenerate marginals, the ε→0 selection principle (the entropic limit picks the most diffusive optimal plan, never a Monge plan when the Schur complement is nonzero), and the non-parametric convergence rates for infinite-dimensional covariances, such as ε^{1−1/(2⌈(n+m)/2⌉)} for integrated Brownian motions.\n\nThe main derivation in Theorem 3.4 is solid: the von Neumann trace inequality reduces the problem to per-eigenvalue scalar optimization, and the resulting Rε = fε(G*M) checks out against the identity case and the orthogonal-marginal example. The bias and W2-rate formulas are internally consistent. There is real content here.\n\nThe soft spot is structural, not technical. The paper's universal claims—arbitrary degenerate A,B, existence of properly aligned Green pairs, the parameterization of all optimal Kantorovich couplings, and the Schur-complement uniqueness criterion—are imported en bloc from the author's own unreviewed preprint [40]. None of those proofs are reproduced. If [40, Thm 2.1] is incomplete, then Theorem 3.4's universality and Theorem 3.9's 'never Monge' selection principle are unsupported. This is not a citation gap; it is a load-bearing dependency. The paper should either prove those results, state them as assumptions, or wait for [40] to clear review.\n\nMinor issues: the O(d) per-ε claim is only for evaluating the regularized cost, not reconstructing the full coupling (which costs O(d²) or O(d³)); the numerics behind Conjecture 3.1 are thin—five trials, no code or seeds; and the text is full of typos. None of these change the verdict.\n\nIf I were the editor, I would send this to review. The referee needs to verify [40]'s Theorem 2.1 and the Schur-complement criterion; if those hold, this is a substantial advance. If not, it is a clean derivation of known results with an unsupported selection principle. Either way it deserves referee time.","headline":"The spectral-shrinkage derivation is clean and the rates are nice, but the universal 'no assumptions' claims and the selection principle rest on the author's own unreviewed preprint [40]—real content, with the universality not yet earned.","tokens_in":25107,"tokens_out":3208,"would_cite":true,"duration_ms":32027,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["49Q22","60B11","47B10","46N30"],"pacs":[],"model":"deepseek-v4-flash","headline":"For any two Gaussian measures, the entropic optimal coupling is uniquely determined by a single spectral filter applied to the aligned correlation operator; as ε→0 it selects the most diffusive optimal plan, never Monge.","keywords":["Entropic Optimal Transport","Gaussian Measures","Bures-Wasserstein Geometry","Covariance Operators","Spectral Shrinkage","Functional Analysis","Schur complement","convergence rates"],"falsifier":"Take A = diag(1,1,0) and B = diag(0,1,1) on R³, build the canonical aligned pair from (6), compute G f_ε(G*M) M* for small ε, and compare the resulting cross-covariance with a Sinkhorn solution of the Gaussian EOT problem; a persistent mismatch beyond numerical tolerance refutes Theorem 3.4. A second check is whether the ε→0 limit equals the N₁₂ = 0 centroid coupling rather than a Monge plan with N₁₂*N₁₂ = B/A.","tokens_in":23907,"feed_emoji":"📉","tokens_out":9964,"duration_ms":85471,"temperature":0.7,"pith_summary":"This paper claims that entropic optimal transport (EOT) between Gaussian measures on a separable Hilbert space—without assuming invertibility or range inclusion of the covariances—has an exact, unique solution given by a universal spectral shrinkage. The optimal cross-covariance is G f_ε(G*M) M*, where G and M are any properly aligned square roots of the two covariances and f_ε(x)=2x/(√(4x²+ε²)+ε) is a single scalar function. This recasts the problem from iterative Sinkhorn computation into a one-time eigendecomposition. As ε→0 the entropic couplings converge in Wasserstein distance to the most diffusive optimal Kantorovich coupling—the centroid of the set of optimal plans—and whenever the Schur complement does not vanish the limit is provably not a Monge solution. The paper also derives closed-form entropic-bias and coupling-distance formulas, recovering the finite-dimensional rate ε log(1/ε) and giving non-parametric rates ε^{1−1/α} for polynomially decaying spectra.","feed_headline":"One universal shrinkage function solves Gaussian entropic transport","feed_subtitle":"For degenerate Gaussian pairs the optimal coupling is f_ε(G*M); at ε→0 it picks the most diffusive plan, never Monge.","key_machinery":"The key object is the properly aligned Green's pair: factorizations G G* = A and M M* = B with G*M ⪰ 0, whose product spectrum is invariant and equals the spectrum of (A^{1/2} B A^{1/2})^{1/2}. The universal spectral shrinkage function f_ε(x) = 2x/(√(4x²+ε²)+ε) acts as the scalar optimizer that decouples the entropic variational problem into independent one-dimensional problems, yielding the closed-form correlation operator R_ε = f_ε(G*M). The generalized Schur complement B/A measures the residual variance in one marginal unexplained by the other and controls uniqueness (B/A = 0) and the failure of Monge selection in the limit.","core_discovery":"The central discovery is Theorem 3.4: for trace-class covariances A and B and any properly aligned Green's operators G and M, the unique Gaussian EOT coupling is N(0, Σ_ε) with cross-covariance G R_ε M*, where R_ε = f_ε(G*M). Because f_ε is a monotone scalar function mapping [0,∞) to [0,1), the entropic solution is a spectral contraction of the correlation operator of the optimal Kantorovich plan. Theorem 3.9 then identifies the ε↓0 limit: the couplings converge in W₂ to the centroid of the convex set of optimal Kantorovich couplings, constructed by setting the free parameter N₁₂ = 0. This implies that in any degenerate regime where the generalized Schur complement B/A is nonzero, the entrop","pith_inferences":["If the spectral-shrinkage formula is taken as the defining structure of a 'Gaussian entropic semigroup', then changing the divergence (for example to an α-divergence) or the reference coupling would correspond to replacing f_ε by another scalar function, and the same spectral decomposition would still yield exact multi-scale paths without re-solving.","The selection of the most diffusive plan at ε→0 is plausibly a general phenomenon in entropic OT beyond Gaussians: in any degenerate problem where the optimal plan set is a convex polytope, entropic regularization may steer the limit to the set's centroid rather than to an extreme Monge point. This is a testable extension, not established by the paper.","The one-step update of Proposition 3.8 offers a practical way to construct proper alignment from arbitrary Cholesky factors; if implemented in high dimensions it would sidestep explicit Schur-complement block decompositions, which matters for operator-valued data.","The non-parametric rates carry an implicit practical warning: for functional data with slow spectral decay (small α), entropic bias decays slowly, so ε should be chosen relative to the spectrum rather than as a universal small constant."],"forward_implications":["For any trace-class A and B, the Gaussian EOT solution is unique for every ε > 0 and independent of the chosen properly aligned Green pair, removing the usual injectivity and range-inclusion assumptions.","As ε → 0 the entropic coupling converges to the most diffusive optimal Kantorovich coupling (the N₁₂ = 0 centroid), so whenever the Schur complement is nonzero the limit is never a Monge plan.","The entropic cost bias has the closed form tr[g_ε((A^{1/2} B A^{1/2})^{1/2})]; in finite rank it scales as ε log(1/ε) with constant equal to the rank, and under polynomial spectral decay s_k = c k^{−α} it scales as ε^{1−1/α}.","A single eigendecomposition of G*M yields exact EOT costs and couplings for every ε > 0—an O(d³) one-time cost followed by O(d) per ε—replacing Sinkhorn iterations and their divergence as ε ↓ 0.","The Wasserstein distance between the entropic coupling and the canonical OT coupling is given explicitly by W₂²(πε, π₀) = 2 tr[A + B − √Q_ε], with operator-based upper bounds expressed through spectral perturbation functions."],"fun_headline_variants":["Gaussian entropic transport is spectral shrinkage","Limit picks most diffusive plan, not Monge","One scalar function shrinks Gaussian transport spectra","Direct algebra replaces Sinkhorn for Gaussian costs","Degenerate limit: entropic transport picks diffusive core"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The degenerate-case conclusions—Theorem 3.9, the never-Monge claim, and the ε→0 selection principle—rest on the parameterization of all optimal Kantorovich couplings imported from the companion preprint [40] (Section 2.2, Theorem 2.1), which is not proved here; the non-parametric rates additionally assume exact polynomial eigenvalue decay s_k = c k^{−α}.","fun_headline_variants_meta":{"raw":{"variants":["Gaussian entropic transport is spectral shrinkage","Limit picks most diffusive plan, not Monge","One scalar function shrinks Gaussian transport spectra","Direct algebra replaces Sinkhorn for Gaussian costs","Degenerate limit: entropic transport picks diffusive core"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000162,"raw_usage":{"total_tokens":1125,"prompt_tokens":841,"completion_tokens":284,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":585,"completion_tokens_details":{"reasoning_tokens":212}},"tokens_in":585,"tokens_out":284,"duration_ms":3971,"temperature":1.0,"reasoning_tokens":212,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T14:45:02.663423+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take A = diag(1,1,0) and B = diag(0,1,1) on R³, build the canonical aligned pair from (6), compute G f_ε(G*M) M* for small ε, and compare the resulting cross-covariance with a Sinkhorn solution of the Gaussian EOT problem; a persistent mismatch beyond numerical tolerance refutes Theorem 3.4. A second check is whether the ε→0 limit equals the N₁₂ = 0 centroid coupling rather than a Monge plan with N₁₂*N₁₂ = B/A.","supporting_citations":[],"review_version":1}