{"id":"0ad9a509-1d0b-4a0f-b0a7-fb30dd6aff4f","arxiv_id":"2505.03202","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The author proves W-entropy dissipation and Shannon entropy power concavity on closed (K,n,N)-super Ricci flows over metric measure spaces, and connects lower-bounded W-entropy to volume non-collapsing on RCD(0,N) spaces.","lead":"This paper extends Perelman's entropy formulas, originally for smooth Ricci flow, to super Ricci flows on metric measure spaces with synthetic curvature bounds. It also states a non-collapsing criterion on RCD(0,N) spaces, but key proofs rely on an unstated Bochner formula.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 4.4/4.5 rests on an unstated distributional Bochner identity in §5.3; without stating/proving it for the regularity class of log u, the main monotonicity claim is unsupported.","rationale":"The stress-test pass confirms the reader's central concern. The paper's Section 3.3 states a distributional Bochner formula (3.17), but Section 5.3 explicitly invokes a different, unnumbered 'Riemannian Bochner formula (??)' as an assumption. This missing formula is load-bearing: it is the step that converts the Γ2-expression in (5.3) into the Hessian/Ric_{N,n}/trace-deviation form in Theorems 4.4 and 4.5, and thus into the non-positivite dissipation bound that drives the W-entropy monotonicity. On RCD spaces, Bochner-type identities are measure-valued and hold for restricted function classes; the theorem hypotheses on u do not guarantee that f=−log u lies in the needed class. This is not merely a missing citation: the claimed theorem scope (all closed RCD spaces) requires the identity to hold in distribution for the low-regularity solution u. Consequently, the central claim is not proven as submitted. The complaint is concrete and internal to the manuscript: the proof text itself flags the unstated formula. I agree with the reader's weakest-assumption identification, and I do not see a reason to change the REJECT verdict. The recommended adjustment is UNCHANGED because the reader's verdict already reflects this gap.","tokens_in":38242,"tokens_out":4511,"duration_ms":46560,"concrete_test":"State the missing identity explicitly in §5.3 with a proof or a precise citation (e.g., a distributional Bochner formula for the time-dependent Witten Laplacian on closed RCD spaces from Li–Zhang, arXiv:2504.01864), and verify it for f = −log u under exactly the hypotheses u∈W^{1,2}∩D(L)∩L∞, Lu∈L∞ of Theorem 4.4/4.5. If the identity can only be proved for f in TestF or for heat-flow-smoothed functions, provide an approximation argument showing the dissipation formula survives in the distributional limit; without such an argument, Theorems 4.4 and 4.5 are unproved in the stated generality.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central monotonicity claim (Theorem 4.5, with Theorem 4.4 as the K=0 case) is proved in §5.3 only 'Under the condition that the Riemannian Bochner formula (??) holds'. The placeholder (??) is never stated or cited. The identity needed is not the paper's Eq. (3.17), which is essentially a definition of Ric(Δ_t) as a measure-valued remainder. What §5.3 requires is a pointwise/measure identity of the form: Γ2(L)(f,f) + (1/t−K) Tr∇²f + (n/4)(1/t−K)² = ||∇²f + (1/2)(1/t−K)g||²_HS + Ric_{N,n}(L)(∇f,∇f) + |(L−Tr∇²)f|²/(N−n), applied with f = log u. This identity is algebraic on smooth spaces given the decomposition Ric∞,n = Ric_{N,n} + (∇φ·∇f)²/(N−n), but on a closed RCD space it presupposes: (i) log u admits a Hessian of finite Hilbert–Schmidt norm, (ii) the measure-valued Γ2 decomposes without a singular part, and (iii) Ric_{N,n}(L) acts as an absolutely continuous quadratic form. The theorem hypotheses on u (u∈W^{1,2}∩D(L)∩L∞, Lu∈L∞) do not by themselves place f=−log u in the usual RCD test-function class TestF for which such Bochner identities are known. Because the proof of the central W-entropy dissipation estimate starts precisely from this unstated formula, the claim 'on every closed RCD space' is not supported as written. The same gap propagates into the dissipation inequality (4.8), the Shannon entropy power theorem, and the LYHP Harnack theorem, which all use the same decomposition; the reader's identification of this as the weakest assumption is accurate.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes to extend Perelman's W-entropy monotonicity, the concavity of Shannon entropy power, and the Li-Yau-Hamilton-Perelman Harnack inequality from smooth super Ricci flows to time-dependent metric measure spaces, in particular to closed (K,n,N)-super Ricci flows defined via a synthetic Bakry-Emery Ricci measure. It also claims applications to the equivalence between volume non-collapsing and lower boundedness of the W-entropy on RCD(0,N) spaces, and to logarithmic Sobolev inequalities. The main new statements are the W-entropy dissipation formulas in Theorems 4.4 and 4.5, with the explicit estimate d/dt W_{N,K}(u) ≤ −(2t/N) ∫ u (L log u + N/(2t) − NK/2)^2 dμ, and the resulting monotonicity under (K,n,N)-super Ricci flow.","tokens_in":38625,"tokens_out":4166,"duration_ms":38903,"significance":"If the main results are correct, the paper would give a genuinely synthetic extension of Perelman's entropy monotonicity, going beyond the static RCD(0,N) results of Kuwada-Li and complementing the recent RCD(K,n,N) work of Li-Zhang and Brena. The explicit dissipation estimate and the Harnack inequalities would be valuable tools for the analysis of time-dependent metric measure spaces. The paper also contains a potentially significant equivalence result for non-collapsing on RCD(0,N) spaces. However, the current manuscript does not establish the central theorems as written because a load-bearing Bochner identity is invoked but never stated or proved, and several supporting estimates and proofs are omitted. These gaps block acceptance and require substantive revision.","major_comments":[{"comment":"The proof begins with 'Under the condition that the Riemannian Bochner formula (??) holds', but the formula is never stated, located, or proved. The identity required is not the paper's Eq. (3.17), which defines Ric∞,n as a measure-valued remainder; rather, the proof needs a pointwise/measure identity expressing Γ2(L)(log u, log u) + (1/t−K) Tr∇² log u + (n/4)(1/t−K)² as the sum of a Hilbert–Schmidt square, a Ric_{N,n}(L) term, and |(L−Tr∇²) log u|²/(N−n). The stated hypotheses on u (u ∈ W^{1,2} ∩ D(L) ∩ L∞, Lu ∈ L∞) do not place −log u in the RCD test-function class TestF, so even the existence of the terms in such a decomposition is not justified. This gap propagates directly to the dissipation estimate (4.8), the Shannon entropy power concavity in Section 6, and the Harnack inequality in Section 7.4, all of which rely on the same decomposition. The central monotonicity claim is therefore unsupported as written.","section":"Section 5.3, proof of Theorems 4.4 and 4.5"},{"comment":"The proof asserts, without proof, that 'Based on the Li-Yau upper bound estimate (8.1), we can prove that ∫ d²(x,y) u(x,y,τ) dμ(y) ≤ C4(N)'. This estimate is then substituted 'into (31)', but no equation (31) exists in the manuscript. The estimate is essential for bounding −∫ v² log v² dμ and hence for deriving the non-collapsing volume lower bound (8.11) from the lower boundedness of W_N. Without a proof or a precise reference for this estimate, the claimed equivalence between volume non-collapsing and lower boundedness of the W-entropy is not established.","section":"Section 8, proof of Theorem 8.1"},{"comment":"Several results that are presented as new theorems of this paper have omitted proofs. Theorem 6.2 says 'the proof has been essentially given by S. Li-Li [37]' with no detail; Theorem 6.3 explicitly omits the second of its two announced proofs; Theorem 9.4 says 'the proof is similar' and leaves the Euler–Lagrange derivation and the monotonicity of μ_K(t) unstated. Since these theorems are part of the paper's claims on metric measure spaces, citing a smooth-manifold proof is not sufficient unless the cited argument is shown to extend verbatim to the stated RCD/super-Ricci-flow setting. This is a completeness gap.","section":"Theorems 6.2, 6.3, and 9.4"},{"comment":"The formula in Theorem 4.5 contains a symbol error: the second integral has 'm − n' in the denominator, although m is not defined in the theorem and the theorem is about the (K,n,N) case; this should be 'N − n'. Additionally, the definition line (4.10) writes 'W_{m,K}(u)' where the theorem is defining W_{N,K}. These typos obscure the already delicate algebraic structure of the entropy formula.","section":"Theorem 4.5, Eq. (4.11)"}],"minor_comments":[{"comment":"Equation (3.17) is called a 'distributional Bochner formula', but as written it is essentially a definition of Ric∞,n as the remainder Γ2(L)(f,f) − ‖∇²f‖²_{HS}. The paper should clearly distinguish this definition from a genuine Bochner identity, which would require proving that Ric∞,n is a measure or a tensor with the expected properties.","section":"Section 3.3, Eq. (3.17)"},{"comment":"The paper assumes the global geometric dimension n of the RCD spaces is constant in t, but this is not automatic for a time-dependent family and is not discussed. The assumption should be stated as an explicit hypothesis of the theorems that use it, not as a general standing assumption.","section":"Section 3.3, Definition 3.7"},{"comment":"The sign and the role of the commutator term [∂_τ, L] log H are not consistent between Lemma 7.11 and Theorem 7.12: the lemma includes a term 2τ[∂_τ,L] log H in the expression for □*w_m, but the theorem's final inequality ≤ 2τP*([∂_τ,L] log H H) seems to drop the negative definite part of WH without stating the necessary inequality WH ≤ 0. This needs clarification.","section":"Theorem 7.12 and Lemma 7.11"},{"comment":"There are numerous typos and infelicities that should be corrected: 'mnaifolds' for 'manifolds', 'rôle' for 'role', 'c Lap lacian' for 'Laplacian', 'Riccatti' for 'Riccati', 'Contempo-rary' for 'Contemporary', 'dimensional' for 'dimensional', and 'dimensioal' for 'dimensional'. Reference [46] is cited as 'arxiv2504.01864' with a missing colon. Theorems 4.6–4.8 should either include the precise 'reasonable growth condition' or state that they are quoted verbatim from [27] and [46].","section":"Throughout"},{"comment":"The remark candidly notes that the results of Section 7.2 were not submitted earlier because the Gaussian lower bound for L ≠ Δ is not true in general. This is a useful caution, but it also means the reader should be told explicitly which of the Harnack statements in Section 7.2 are actually proved under the stated hypotheses and which are conditional on unverified assumptions.","section":"Section 7.2, Remark 7.8"}],"recommendation":"major_revision","confidential_remarks":"The manuscript leans heavily on the companion paper [46] and on the author's own previous works. The editor may wish to verify that the distributional Bochner identity for the time-dependent Witten Laplacian on RCD spaces invoked in §5.3 is indeed proved in [46] or in the RCD literature; otherwise the central monotonicity theorem is conditional on an unstated and possibly unavailable ingredient. The paper's scope is suitable for the journal, but the missing formula and omitted estimates are significant enough to require a genuine revision rather than editorial fixes."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper is a reasonable next step: it sets up (K,n,N)-super Ricci flows on mm spaces and writes down the Perelman W-entropy in that setting, with the static RCD case already treated by Kuwada–Li and Li–Zhang. The time-dependent statements are new as formulations, and the applications (LYHP Harnack, non-collapsing vs lower-bounded W-entropy, log-Sobolev) are natural. That is real work, and the literature is handled honestly.\n\nThe problem is the proof of the central theorem. Theorems 4.4 and 4.5 are derived in §5.3 by invoking a \"Riemannian Bochner formula (??)\" that is never stated, cited, or proved. This is not a cosmetic gap. The identity needed is stronger than Eq. (3.17); it has to decompose Γ2(L)(log u, log u) into a squared Hessian term plus Ric_{N,n} plus a (L−Tr ∇²)²/(N−n) term. On a smooth space this is algebraic, but on a general closed RCD space it requires log u to have a Hessian of finite Hilbert–Schmidt norm and the measure-valued Γ2 to have no singular part. The hypotheses u∈W^{1,2}∩D(L)∩L∞ with Lu∈L∞ do not put f=−log u in the test-function class where such identities are known. Until that is supplied, the monotonicity claim \"on every closed RCD space\" is unsupported.\n\nOther soft spots are smaller but real. Theorem 8.1's proof mentions an estimate ∫ d²u ≤ C4(N) and refers to equation (31), which does not exist. The proofs of Theorems 6.2, 6.3, and 9.4 are omitted. These are probably fixable, but they are not in the paper.\n\nThe paper deserves a serious referee rather than a desk reject: the definitions are sensible, the missing Bochner identity looks repairable, and the program is coherent. But in its current form the central claim is not proven. I would recommend major revision, not acceptance. If I were the editor I would send it out, and if asked I'd advise rejecting the present version and inviting a resubmission that states and proves the Bochner identity—or narrows the hypotheses to a class where it holds—and fills the gaps.","headline":"Perelman's W-entropy on synthetic super Ricci flows: the program is right, but the central theorem currently rests on an unstated Bochner identity and needs a major revision before the main claim is supported.","tokens_in":39285,"tokens_out":3242,"would_cite":false,"duration_ms":31795,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C23","53C21","60J60","60H30"],"pacs":[],"model":"deepseek-v4-flash","headline":"Perelman's W-entropy monotonicity is extended to super Ricci flows on metric measure spaces.","keywords":["Li-Yau-Hamilton-Perelman Harnack inequality","Log-Sobolev inequality","Perelman's W-entropy","Shannon entropy power","super Ricci flows","metric measure spaces","volume non-local collapsing property"],"falsifier":"Construct a closed $(0,n,N)$-super Ricci flow on a metric measure space that satisfies the definition but for which the distributional Bochner identity $\\Gamma_2(L_t)(f,f) = \\|\\mathrm{Hess}_t f\\|^2_{\\mathrm{HS}} + \\mathrm{Ric}_{N,n}(L_t)(\\nabla f,\\nabla f)$ fails, for example a space with a singular potential $\\varphi_t$ outside the assumed differentiability class; then compute $dW_N/dt$ directly and look for a time where it becomes positive, which would refute Theorem 4.4.","tokens_in":37931,"feed_emoji":"📉","tokens_out":6180,"duration_ms":54922,"temperature":0.7,"pith_summary":"This paper establishes that Perelman's W-entropy, a functional over solutions of the conjugate heat equation that is monotone on smooth Ricci flows, remains non-increasing on 'super Ricci flows' on metric measure spaces, i.e., spaces with synthetic Ricci curvature lower bounds. The main theorems give explicit dissipation formulas for the W-entropy on closed $(K,n,N)$-super Ricci flows and derive the Shannon entropy power concavity, a Li-Yau-Hamilton-Perelman differential Harnack inequality, an equivalence between volume non-local collapsing and lower boundedness of the W-entropy, and optimal log-Sobolev inequalities. If correct, this extends a tool central to the analysis of Ricci flow singularities to the non-smooth, synthetic setting.","feed_headline":"Entropy monotonicity proven for metric-measure super Ricci flows","feed_subtitle":"Perelman's W-entropy never increases along the heat flow on closed synthetic spaces with Ricci lower bounds.","key_machinery":"The central object is the W-entropy functional $W_{N,K}(u,t) = \\int_X\\big[t|\\nabla f|^2 + f - N(1+\\frac{Kt}{2})^2\\big]u\\,d\\mu$, where $u = e^{-f}/(4\\pi t)^{N/2}$ solves the heat equation of the time-dependent Witten Laplacian $L_t = \\Delta_t - \\nabla\\varphi_t\\cdot\\nabla$. The argument is carried by the conjugate heat equation $\\frac{d}{dt}(e^{-\\varphi_t}dm_t)=0$, the definition of a $(K,n,N)$-super Ricci flow via the $N$-dimensional Bakry-Emery Ricci curvature measure, and a distributional Bochner identity $\\Gamma_2(L_t)(f,f) = \\|\\mathrm{Hess}_t f\\|^2_{\\mathrm{HS}} + \\mathrm{Ric}_{N,n}(L_t)(\\nabla f,\\nabla f)$. These combine to produce the dissipation formulas (4.7) and (4.11).","core_discovery":"The central claim is that on every closed $(K,n,N)$-super Ricci flow on metric measure spaces satisfying the conjugate heat equation, the W-entropy $W_{N,K}(u,t)$ is non-increasing in time, with the explicit dissipation estimate $$\\frac{d}{dt}W_{N,K}(u) \\le -\\frac{2t}{N}\\int_X u\\Big(L\\log u + \\frac{N}{2t} - \\frac{NK}{2}\\Big)^2 d\\mu$$ (Theorems 4.4 and 4.5). This extends Perelman's entropy monotonicity from smooth Ricci flows to synthetic spaces. The paper also proves the corresponding concavity of the Shannon entropy power, a Li-Yau-Hamilton-Perelman Harnack inequality, and the equivalence between the volume non-collapsing property and lower boundedness of the W-entropy on RCD$(0,N)$ spaces.","pith_inferences":["If the missing Bochner identity is supplied, the same dissipation formulas imply analogous monotonicity for $(K,\\infty)$-super Ricci flows, where the finite-$N$ terms disappear.","The equivalence between non-collapsing and lower bounded W-entropy suggests a quantitative stability statement: a lower bound on $W$ with a given constant $A$ implies explicit volume growth with constant $e^{-A}$, as in (8.11).","The entropy power concavity on synthetic spaces could be used to prove information-theoretic inequalities, such as Costa's entropy power inequality, for heat semigroups on RCD spaces.","The regularity issue noted in Remark 9.5 for the extremal of the log-Sobolev functional may block the Euler-Lagrange characterization on non-smooth spaces; a nonsmooth counterexample would pinpoint the limit of Theorem 9.4."],"forward_implications":["The W-entropy monotonicity provides a Lyapunov function for the conjugate heat flow on synthetic spaces, so Perelman-style non-collapsing arguments can be ported to RCD spaces.","The Shannon entropy power $e^{2H/N}$ is concave (or $(-2K)$-concave) along the heat flow on closed $(0,n,N)$ (or $(K,n,N)$) super Ricci flows, giving a sharp Fisher information bound $I(u(t)) \\le N/(2t)$.","The Li-Yau-Hamilton-Perelman Harnack inequality holds for the fundamental solution on such spaces, yielding differential Harnack estimates.","On RCD$(0,N)$ spaces, volume non-local collapsing is equivalent to lower boundedness of the W-entropy, and the limit of the W-entropy equals the logarithm of the volume ratio constant."],"supporting_citations":[{"why":"introduces the W-entropy and its monotonicity for closed Ricci flows, the result being extended here.","marker":"[51]"},{"why":"proves the W-entropy dissipation formula on smooth $(K,m)$-super Ricci flows, the smooth prototype for Theorems 4.4 and 4.5.","marker":"[30]"},{"why":"proves W-entropy monotonicity and rigidity on RCD$(0,N)$ spaces, the static synthetic case generalized here.","marker":"[27]"},{"why":"proves the W-entropy formula on RCD$(K,n,N)$ spaces, supplying the method and Bochner-type identities used in the time-dependent setting.","marker":"[46]"},{"why":"defines super Ricci flows for metric measure spaces and supplies the $(K,N)$-super Ricci flow notion used in Theorem 4.3.","marker":"[55]"},{"why":"establishes the equivalence between dynamic $(K,N)$-convexity and super-$N$-Ricci flow conditions, background for Theorem 4.3.","marker":"[29]"},{"why":"derives the W-entropy formula for the Witten Laplacian on weighted manifolds, the technical template for the proofs.","marker":"[42]"},{"why":"introduced the W-entropy for the linear heat equation and the volume-growth equivalence, which Section 8 adapts.","marker":"[49]"}],"fun_headline_variants":["W-entropy monotonicity for synthetic Ricci flows","Perelman's entropy holds on metric-measure spaces","Super Ricci flows: entropy never increases","Shannon entropy power concavity on super Ricci","Volume non-collapsing tied to W-entropy bound"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proofs of Theorems 4.4 and 4.5 require a distributional Bochner formula for the time-dependent Witten Laplacian on closed RCD spaces, which the paper invokes as 'the Riemannian Bochner formula (??)' without stating or citing it; if that identity is not valid on the spaces covered, the W-entropy monotonicity is unsupported.","fun_headline_variants_meta":{"raw":{"variants":["W-entropy monotonicity for synthetic Ricci flows","Perelman's entropy holds on metric-measure spaces","Super Ricci flows: entropy never increases","Shannon entropy power concavity on super Ricci","Volume non-collapsing tied to W-entropy bound"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000671,"raw_usage":{"total_tokens":3003,"prompt_tokens":836,"completion_tokens":2167,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":452,"completion_tokens_details":{"reasoning_tokens":2094}},"tokens_in":452,"tokens_out":2167,"duration_ms":15975,"temperature":1.0,"reasoning_tokens":2094,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:57:52.398044+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a closed $(0,n,N)$-super Ricci flow on a metric measure space that satisfies the definition but for which the distributional Bochner identity $\\Gamma_2(L_t)(f,f) = \\|\\mathrm{Hess}_t f\\|^2_{\\mathrm{HS}} + \\mathrm{Ric}_{N,n}(L_t)(\\nabla f,\\nabla f)$ fails, for example a space with a singular potential $\\varphi_t$ outside the assumed differentiability class; then compute $dW_N/dt$ directly and look for a time where it becomes positive, which would refute Theorem 4.4.","supporting_citations":[{"cited_title":", http://arXiv.org/abs/maths0211159","cited_arxiv_id":null,"evidence_quote":"introduces the W-entropy and its monotonicity for closed Ricci flows, the result being extended here."},{"cited_title":"Pacific J","cited_arxiv_id":null,"evidence_quote":"proves the W-entropy dissipation formula on smooth $(K,m)$-super Ricci flows, the smooth prototype for Theorems 4.4 and 4.5."},{"cited_title":"Manuscripta Math","cited_arxiv_id":null,"evidence_quote":"proves W-entropy monotonicity and rigidity on RCD$(0,N)$ spaces, the static synthetic case generalized here."},{"cited_title":"On the $W$-entropy and Shannon entropy power on RCD$(K, N)$ and RCD$(K, n, N)$ spaces","cited_arxiv_id":"2504.01864","evidence_quote":"proves the W-entropy formula on RCD$(K,n,N)$ spaces, supplying the method and Bochner-type identities used in the time-dependent setting."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"defines super Ricci flows for metric measure spaces and supplies the $(K,N)$-super Ricci flow notion used in Theorem 4.3."},{"cited_title":"Pure and Appl","cited_arxiv_id":null,"evidence_quote":"establishes the equivalence between dynamic $(K,N)$-convexity and super-$N$-Ricci flow conditions, background for Theorem 4.3."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"derives the W-entropy formula for the Witten Laplacian on weighted manifolds, the technical template for the proofs."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"introduced the W-entropy for the linear heat equation and the volume-growth equivalence, which Section 8 adapts."}],"review_version":1}