REVIEW 1 major objections 6 minor 38 references
A heat-flow inequality proves the optimal mean-width bound for isotropic convex bodies, M*(K) ≤ C√(n log n), and sharpens the MM* product estimate to C log n.
Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →
T0 review · deepseek-v4-flash
2026-08-03 06:43 UTC pith:VTOST4CW
load-bearing objection Clean new heat-flow inequality; main theorem is conditional on an external κ_n bound the paper doesn't fully prove — still worth refereeing. the 1 major comments →
Optimal MM^* bounds for convex bodies
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
For every centered convex body K in R^n, the paper proves (E h_K(G))^2 ≤ 2n E h_K(X_K), where G is a standard Gaussian vector, X_K is uniform on K, and h_K is the support function. This heat-flow inequality requires no isotropy and is the load-bearing new step. Applying it to an isotropic body and combining with the gauge-order comparison E h_K(X_K) ≤ C√(log n) E h_K(G) (which itself now holds with a universal constant thanks to the dimension-free bound on κ_n) yields M*(K) ≤ C√(n log n). The same combination with the mean-norm estimate M(K) ≤ C√(log n / n) gives the product bound ℓ(K) ≤ C log n, extending Pisier's MM* estimate to arbitrary convex bodies in isotropic position.
What carries the argument
The central identity is the heat-flow inequality of Theorem 1.3: (E h_K(G))^2 ≤ 2n E h_K(X_K), proved by running a Brownian motion from the uniform measure on K and using the fact that the heat semigroup of the support function has gradient in K, so the boundary term is controlled by the support function of the outward normal. The other ingredient is the gauge comparison theorem, which says for any gauge ∥·∥ and isotropic log-concave X, E∥X∥ is comparable to E∥G∥ up to √(log n) factors; its right-hand side now has a universal constant because the third-moment parameter κ_n satisfies κ_n ≤ 4, a dimension-free bound imported from an external preprint and only sketched here.
Load-bearing premise
The deduction of the mean-width theorem relies on the external dimension-free bound on the third-moment parameter κ_n (equivalently, the anisotropic thin-shell estimate Var(|X|^2) ≤ 8 Tr(Σ^2)), which the paper quotes from another preprint and only sketches; if that bound fails, the main theorem does not follow from this proof.
What would settle it
Construct a sequence of isotropic log-concave vectors X_n for which sup over gauges ∥·∥ of E∥X_n∥ / (√(log n) E∥G∥) tends to infinity, or directly exhibit a centered log-concave vector with covariance Σ such that Var(|X|^2) > 8 Tr(Σ^2) (or κ_n > 4). Either would falsify the imported dimension-free estimate that the proof needs.
If this is right
- The mean-width of any isotropic convex body is O(√(n log n)), and this is optimal since the isotropic cross-polytope attains the order.
- The product M(K)M*(K) in isotropic position is O(log n), sharp even in the origin-symmetric class and even if one optimizes over affine positions.
- Pisier's MM* estimate now holds for non-symmetric convex bodies, with the position chosen to be isotropic rather than a special linear image.
- Banach–Mazur distances between convex bodies are bounded by C n log n, improving the previously known polylogarithmic bounds.
- The flatness constant satisfies Flt(n) ≤ C n log n, improving the earlier O(n log^2 n) bound.
Where Pith is reading between the lines
- The heat-flow inequality itself is position-free and self-contained; it may hold with better constants or extend to other homogeneous convex functions, offering a route to mean-width control without stochastic localization.
- The proof of Theorem 1.1 is conditional on the imported dimension-free bound κ_n ≤ 16; if that bound fails, the mean-width theorem would still hold by the heat-flow inequality alone only up to an unknown n^ε factor, not the sharp √(log n).
- The sharpness example Q_m × P_m suggests that to prove Conjecture 1.8 (√(log n) MM* for symmetric bodies) one must leave the isotropic position; the paper's own product example shows the isotropic position can force the larger log n factor.
- The anisotropic thin-shell estimate Var(|X|^2) ≤ 8 Tr(Σ^2) may be of independent interest in high-dimensional probability, and the paper sketches how it also yields the KLS bound Ψ_n ≤ C (log n)^{1/4}.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proves the optimal mean-width bound M*(K) ≤ C√(n log n) for isotropic convex bodies in R^n (Theorem 1.1). The main new ingredient is Theorem 1.3, a heat-flow inequality showing (E h_K(G))^2 ≤ 2n E h_K(X_K). Combining this with the gauge comparison of Eldan–Lehec and Bizeul–Klartag (Theorem 1.2), whose statements become dimension-free assuming a bound κ_n^2 ≤16, yields the mean-width estimate. The paper also derives the MM* product bound M(K)M*(K) ≤ C log n (Corollary 1.5), discusses sharpness and affine optimality, and records applications to Banach–Mazur distances and the flatness constant.
Significance. The mean-width estimate matches the lower bound from the cross-polytope and improves the previous O(√n log^2 n) bound of E. Milman. The heat-flow argument in Theorem 1.3 is elegant, self-contained, and of independent interest. If the external κ_n bound is valid, the paper resolves the mean-width problem in the isotropic position and gives the first MM* estimate for non-symmetric bodies in an explicit position, with optimal log n order. The optimality examples and applications are well chosen. The main weakness is that the proof of Theorem 1.1 is not self-contained: it relies on the dimension-free bound κ_n^2 ≤16 from the unpublished preprint [31], and Section 2 is explicitly expository.
major comments (1)
- [Section 2 / Theorem 1.1] The proof of Theorem 1.1 is conditional on the external bound κ_n^2 ≤16 (Corollary 2.7) from [31]. Section 2 is explicitly 'purely expository', and the anisotropic thin-shell estimate (Theorem 2.5) is proved only under smoothness assumptions, with the approximation step delegated to [11]. Since Theorem 1.2 and hence Theorem 1.1 require this dimension-free bound, the central claim is not supported by the present manuscript alone. Please either include a complete proof of the κ_n bound (including the approximation argument) in an appendix, or explicitly state Theorems 1.1, 1.2 and Corollary 1.5 as conditional on [31]. As written, a failure of [31] would invalidate the main theorem.
minor comments (6)
- [Section 2, Corollary 2.7] The inequality ||B||² = E(⟨X,θ⟩⟨BX,X⟩) ≤ Var(⟨BX,X⟩)^{1/2} is only true after centering: use E⟨X,θ⟩=0 to write the left side as E(⟨X,θ⟩(⟨BX,X⟩−Tr(B))). Please add this justification.
- [Section 2, Theorem 2.5] The passage from the regular to the general case of centered log-concave vectors is delegated to [11]; a brief explanation or a precise reference to the approximation argument would help the reader.
- [Section 3.1, Theorem 1.3] The commutation ∇P_s h_K = P_s(∇h_K) and the differentiation under the integral sign are used without comment; a sentence on the regularization argument would improve clarity.
- [Section 3.3] In the product example K=Q_m×P_m, the intermediate value E h_{P_m}(G_2) ≃ m√(log m) is later divided by E|G|; consider making the normalization explicit.
- [Introduction] Typo: 'discusseion' should be 'discussion' in the paragraph before Notation.
- [References] Reference [31] is an arXiv preprint; if the manuscript is intended for publication, the status of this reference should be updated or the dependence should be addressed as in Major Comment 1.
Circularity Check
No circularity: the central heat-flow inequality (Theorem 1.3) is proved from scratch, and the gauge comparison and κ_n bound are independent external inputs.
full rationale
The derivation chain is not circular. Theorem 1.3, the main new ingredient, is proved by a direct heat-flow argument (§3.1) and does not assume Theorem 1.1, Theorem 1.2, or any fitted value. The deduction of Theorem 1.1 in §3.2 combines Theorem 1.3 with the right-hand gauge comparison of Theorem 1.2, which is attributed to Eldan and Lehec [15]; the left-hand comparison from [7] is not used in that deduction. The dimension-free bound κ_n^2 ≤ 16 is imported from the external preprint [31] and only sketched in §2, which the paper itself labels 'purely expository and contains no original content.' This is a genuine conditionality—if [31] were false, Theorem 1.1 would lack support—but it is not circularity: [31] is not derived from the paper's own conclusions, and Theorem 1.3 is internally proved. The use of the author's own [7] for the mean-norm half of Corollary 1.5 and of [6] for the slicing step is also not circular: these are prior independent results with their own stated assumptions, not restatements of the present theorem. No parameter is fitted to the predicted quantities, and no uniqueness or ansatz is imported from a self-citation to force the main result. The paper is therefore best assessed as non-circular, with its reliance on external results noted as a limitation rather than a circular step.
Axiom & Free-Parameter Ledger
axioms (8)
- domain assumption Right-hand gauge comparison for isotropic log-concave vectors: E||X|| <= C sqrt(log n) E||G|| for every gauge (Theorem 1.2, from Eldan-Lehec [15] with kappa_n)
- domain assumption Dimension-free third-moment bound kappa_n^2 <= 16 (Eq. (14), from [31])
- domain assumption Sharp thin-shell estimate Var(|X|^2) <= 8n for isotropic log-concave X (Chen-Klartag [11])
- domain assumption Slicing problem solution: for volume-normalized K, the isotropic scaling lambda_K is of order sqrt(n) ([27],[6])
- domain assumption Giannopoulos-Milman characterization of minimal mean-width position (Eq. (51))
- domain assumption Banaszczyk-Litvak-Pajor-Szarek lower bound ell(Delta) >= c log n for simplices
- standard math Heat semigroup, divergence theorem, and support function properties used in the proof of Theorem 1.3
- standard math Brascamp-Lieb inequality and moment-measure theorem in Section 2
read the original abstract
Let $K\subset\R^n$ be a convex body in isotropic position. We prove the optimal mean-width estimate \[ M^*(K)\leq C\sqrt{n\log n}. \] The main new ingredient is a geometric inequality relating the Gaussian mean of the support function to its mean under the uniform measure on $K$, obtained through a heat-flow argument. Combined with the Gaussian-log-concave comparison of Eldan and Lehec and the newly available dimension-free bound on the third-moment parameter $\kappa_n$, this yields the result. The boundedness of $\kappa_n$ also makes the mean-norm estimate of Bizeul and Klartag sharp. Combining both estimates yields \[ M(K)M^*(K)\leq C\log n, \] extending Pisier's $MM^*$ estimate to non-symmetric convex bodies and to the isotropic position.
Reference graph
Works this paper leans on
-
[1]
Alonso-Guti´ errez and J
D. Alonso-Guti´ errez and J. Bastero,The variance conjecture on some polytopes, in Asymptotic Geometric Analysis, Fields Inst. Commun. 68, Springer, New York, 2013, 1-20
2013
-
[2]
Anttila, K
M. Anttila, K. Ball, and I. Perissinaki,The central limit problem for convex bodies, Trans. Amer. Math. Soc.355(2003), no. 12, 4723-4735
2003
-
[3]
Artstein-Avidan, A
S. Artstein-Avidan, A. Giannopoulos, and V. D. Milman,Asymptotic Geometric Anal- ysis, Part I, Mathematical Surveys and Monographs, Vol. 202, American Mathematical Society, Providence, RI, 2015
2015
-
[4]
Banaszczyk, A
W. Banaszczyk, A. E. Litvak, A. Pajor, and S. J. Szarek,The flatness theorem for nonsymmetric convex bodies via the local theory of Banach spaces, Math. Oper. Res. 24(1999), no. 3, 728-750
1999
-
[5]
Barthe and B
F. Barthe and B. Klartag,Spectral gaps, symmetries and log-concave perturbations, Bull. Hellenic Math. Soc.64(2020), 1-31
2020
-
[6]
Bizeul,The slicing conjecture via small ball estimates, arXiv:2501.06854, 2025
P. Bizeul,The slicing conjecture via small ball estimates, arXiv:2501.06854, 2025
Pith/arXiv arXiv 2025
-
[7]
P. Bizeul and B. Klartag,Distances between non-symmetric convex bodies: optimal bounds up to polylog, arXiv:2510.20511v3, 2026. 18
Pith/arXiv arXiv 2026
-
[8]
S. G. Bobkov and A. Koldobsky,On the central limit property of convex bodies, in Ge- ometric Aspects of Functional Analysis (2001-02), Lecture Notes in Math., Vol. 1807, Springer, 2003, 44-52
2001
-
[9]
H. J. Brascamp and E. H. Lieb,On extensions of the Brunn-Minkowski and Pr´ ekopa- Leindler theorems, including inequalities for log-concave functions, J. Funct. Anal.22 (1976), no. 4, 366-389
1976
-
[10]
Brazitikos, A
S. Brazitikos, A. Giannopoulos, P. Valettas, and B.-H. Vritsiou,Geometry of Isotropic Convex Bodies, Mathematical Surveys and Monographs, Vol. 196, American Mathe- matical Society, Providence, RI, 2014
2014
-
[11]
Y. Chen and B. Klartag,Digesting the proof of the sharp thin-shell inequality, arXiv:2607.23307, 2026
Pith/arXiv arXiv 2026
-
[12]
Cordero-Erausquin and B
D. Cordero-Erausquin and B. Klartag,Moment measures, J. Funct. Anal.268(2015), no. 12, 3834-3866
2015
-
[13]
T. A. Courtade, M. Fathi, and A. Pananjady,Existence of Stein kernels under a spectral gap, and discrepancy bounds, Ann. Inst. Henri Poincar´ e Probab. Stat.55 (2019), no. 2, 777-790
2019
-
[14]
Eldan,Thin shell implies spectral gap up to polylog via a stochastic localization scheme, Geom
R. Eldan,Thin shell implies spectral gap up to polylog via a stochastic localization scheme, Geom. Funct. Anal.23(2013), no. 2, 532-569
2013
-
[15]
Eldan and J
R. Eldan and J. Lehec,Bounding the norm of a log-concave vector via thin-shell es- timates, in Geometric Aspects of Functional Analysis: Israel Seminar (GAF A) 2011- 2013, Springer, 2014, 107-122
2011
-
[16]
Fathi,Stein kernels and moment maps, Ann
M. Fathi,Stein kernels and moment maps, Ann. Probab.47(2019), no. 4, 2172-2185
2019
-
[17]
A. A. Giannopoulos and V. D. Milman,Extremal problems and isotropic positions of convex bodies, Israel J. Math.117(2000), 29-60
2000
-
[18]
Giannopoulos and E
A. Giannopoulos and E. Milman,M-estimates for isotropic convex bodies and theirL q- centroid bodies, inGeometric Aspects of Functional Analysis, Lecture Notes in Math. 2116, Springer, Cham, 2014, 159–182
2014
-
[19]
Giannopoulos, P
A. Giannopoulos, P. Stavrakakis, A. Tsolomitis, and B.-H. Vritsiou,Geometry of the Lq-centroid bodies of an isotropic log-concave measure, Trans. Amer. Math. Soc.367 (2015), no. 7, 4569–4593
2015
-
[20]
Guan,A note on Bourgain ’s slicing problem, arXiv:2412.09075, 2024
Q. Guan,A note on Bourgain ’s slicing problem, arXiv:2412.09075, 2024
Pith/arXiv arXiv 2024
-
[21]
Kannan, L
R. Kannan, L. Lov´ asz, and M. Simonovits,Isoperimetric problems for convex bodies and a localization lemma, Discrete Comput. Geom.13(1995), 541-559
1995
-
[22]
Klartag,A Berry-Esseen type inequality for convex bodies with an unconditional basis, Probab
B. Klartag,A Berry-Esseen type inequality for convex bodies with an unconditional basis, Probab. Theory Related Fields145(2009), no. 1-2, 1-33. 19
2009
-
[23]
Klartag,Logarithmic bounds for isoperimetry and slices of convex sets, Ars Inve- niendi Analytica, Paper No
B. Klartag,Logarithmic bounds for isoperimetry and slices of convex sets, Ars Inve- niendi Analytica, Paper No. 4 (2023), 1-17
2023
-
[24]
Klartag,Logarithmically-concave moment measures I, in Geometric Aspects of Functional Analysis, Lecture Notes in Math., Vol
B. Klartag,Logarithmically-concave moment measures I, in Geometric Aspects of Functional Analysis, Lecture Notes in Math., Vol. 2116, Springer, Cham, 2014, 231- 260
2014
-
[25]
B. Klartag and J. Lehec,Thin-shell bounds via parallel coupling, arXiv:2507.15495v2, 2026
arXiv 2026
-
[26]
Klartag and J
B. Klartag and J. Lehec,Bourgain ’s slicing problem and KLS isoperimetry up to poly- log, Geom. Funct. Anal.32(2022), no. 5, 1134-1159
2022
-
[27]
Klartag and J
B. Klartag and J. Lehec,Affirmative resolution of Bourgain ’s slicing problem using Guan ’s bound, Geom. Funct. Anal.35(2025), 1147-1168
2025
-
[28]
Klartag and J
B. Klartag and J. Lehec,Isoperimetric inequalities in high-dimensional convex sets, Bull. Amer. Math. Soc.62(2025), 575-642
2025
-
[29]
Hartzoulaki,Probabilistic methods in the theory of convex bodies, Ph.D
M. Hartzoulaki,Probabilistic methods in the theory of convex bodies, Ph.D. thesis, University of Crete, March 2003
2003
-
[30]
Y. T. Lee and S. Vempala,Eldan ’s stochastic localization and the KLS conjecture: isoperimetry, concentration and mixing, Ann. of Math. (2)199(2024), no. 3, 1043- 1092
2024
-
[31]
Letwin,The KLS constant isO(log 1/4 n), arXiv:2607.24164, 2026
B. Letwin,The KLS constant isO(log 1/4 n), arXiv:2607.24164, 2026
Pith/arXiv arXiv 2026
-
[32]
Milman,On the mean-width of isotropic convex bodies and their associatedL p- centroid bodies, Int
E. Milman,On the mean-width of isotropic convex bodies and their associatedL p- centroid bodies, Int. Math. Res. Not. IMRN (2015), no. 11, 3408-3423
2015
-
[33]
Paouris,Concentration of mass on convex bodies, Geom
G. Paouris,Concentration of mass on convex bodies, Geom. Funct. Anal.16(2006), no. 5, 1021–1049
2006
-
[34]
Pisier,The Volume of Convex Bodies and Banach Space Geometry, Cambridge Tracts in Mathematics, Vol
G. Pisier,The Volume of Convex Bodies and Banach Space Geometry, Cambridge Tracts in Mathematics, Vol. 94, Cambridge University Press, Cambridge, 1989
1989
-
[35]
Pivovarov,On the volume of caps and bounding the mean-width of an isotropic convex body, Math
P. Pivovarov,On the volume of caps and bounding the mean-width of an isotropic convex body, Math. Proc. Cambridge Philos. Soc.149(2010), no. 2, 317–331
2010
-
[36]
V. Reis and T. Rothvoss,The subspace flatness conjecture and faster integer program- ming, arXiv:2303.14605v5, 2026
arXiv 2026
-
[37]
Rudelson,Distances between non-symmetric convex bodies and theM M ∗-estimate, Positivity4(2000), no
M. Rudelson,Distances between non-symmetric convex bodies and theM M ∗-estimate, Positivity4(2000), no. 2, 161-178
2000
-
[38]
Vritsiou,Regular ellipsoids and a Blaschke–Santal´ o-type inequality for projec- tions of non-symmetric convex bodies, J
B.-H. Vritsiou,Regular ellipsoids and a Blaschke–Santal´ o-type inequality for projec- tions of non-symmetric convex bodies, J. Funct. Anal.286(2024), no. 11, Paper No. 110414, 40 pp. 20
2024
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.