Heat kernel geometry and Gromov's volume growth conjecture

Jian Ge

Heat kernel geometry and Gromov's volume growth conjecture

T0 review · 0 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Heat flow proves Gromov's volume growth bound: balls grow at most R^(n-2)

desk verdict A serious, largely self-contained proof of the uniform form of Gromov's volume-growth conjecture; I found no internal error, but Lemma B.1 deserves independent audit before calling it settled. read the letter →

arxiv 2608.13553 v1 pith:ETO37DRX submitted 2026-08-13 math.DG

Jian Ge This is my paper · ORCID

classification math.DG MSC 53C2153C2358J35

keywords GromovvolumegrowthconjecturenonnegativeRiccicurvaturepositivescalarheatkernelFisherinformationmetricNashentropyGaussianheat-kernelestimate

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper claims to settle Gromov's 1986 volume growth question in its uniform form: every complete, connected, noncompact $n$-dimensional Riemannian manifold, $n\ge 3$, with nonnegative Ricci curvature and scalar curvature at least $1$ satisfies $\operatorname{Vol}_g(B(p,R))\le C_n R^{n-2}$ for all centers $p$ and radii $R>0$. The exponent $n-2$ is sharp, because a round $2$-sphere of radius $a$ times Euclidean $\mathbb{R}^{n-2}$ has exactly this growth and satisfies both curvature hypotheses. The proof introduces a heat-flow picture in which the conjectured loss of two dimensions appears as a trace defect in the heat-kernel Fisher metric, and shows that this defect is fed by at least two units of scalar-curvature production. A resulting Nash-entropy decay, combined with the standard Gaussian heat-kernel estimate, converts the curvature information into the volume bound.

What carries the argument

The carrying object is the heat-kernel Fisher metric $G_x(t)$, the trace defect $E_x(t)$ it defines, and the centered production source $Q_x(t)=4t^2D_x(t)+|g_x-G_x(t)|^2+2t\langle \mathrm{Ric}_x,G_x(t)\rangle$. The metric bounds $0\le G\le g$, proved from reverse Poincaré and logarithmic Sobolev inequalities, make $E$ nonnegative. The two source identities $\Box_x E=Q/2$ and $\Box_x(-S)=E/t^2$ are pointwise equations in the pole variable; the central analytic step is a spectral-cutoff and weighted Weitzenböck estimate showing that any region where $Q<2-\varepsilon$ carries a distinguished lowest covariance eigenline with definite Ricci curvature, so its heat-averaged mass is small. This gives the two-unit lower bound $P_s(Q_\bullet(t))(z)\ge 2-C_n(t^{-1}+s^{-1})^{1/3}$, and a minimal heat-potential comparison propagates the two units into the entropy decay.

What would settle it

The claim would be settled false by any complete noncompact $n$-manifold with $\mathrm{Ric}\ge 0$, $\mathrm{Scal}\ge 1$, and $\sup_{p,R} \operatorname{Vol}_g(B(p,R))/R^{n-2}=\infty$; a concrete candidate would be a sequence of manifolds with spherical factors whose radii grow slowly, where one could check numerically whether the heat-averaged source $P_s(Q_\bullet(t))(z)$ dips below $2$ at large $t$ and small $s$.

Watch

Extended reading notes

Core claim

The central claim is Theorem 1.1: under $\operatorname{Ric}_g\ge 0$ and $\operatorname{Scal}_g\ge 1$ on a complete noncompact $n$-manifold, $n\ge 3$, there is a dimension-only constant $C_n$ with $\operatorname{Vol}_g(B(p,R))\le C_n R^{n-2}$ for every $p$ and every $R>0$. The sharp model is the product $S^2_a\times \mathbb{R}^{n-2}$, whose balls grow like $4\pi a^2\omega_{n-2}R^{n-2}$. The paper's route is new: it regards the heat kernel as a map into probability densities, studies the normalized Fisher-information metric $G_x(t)=2t\int_M d_x h\otimes d_x h\,d\mu_{x,t}$, and measures how many directions the heat flow sees at scale $\sqrt{t}$. The trace defect $E_x(t)=\frac{t}{2}(n-\operatorname{tr}_g G_x(t))$ vanishes on Euclidean space and approaches $2$ on the sharp model, and the proof shows that its source $Q$ carries a heat-averaged lower bound of two units. Propagating this source through the equations $\Box_x E=Q/2$ and $\Box_x(-S)=E/t^2$ yields the entropy decay $S_x(T)\le -\log T+C_n$, which the Gaussian heat-kernel bound converts into the claimed volume growth.

Load-bearing premise

The proof depends on the uniform pole-variable heat-kernel jet estimates in Lemma B.1, which assert that every pole-space-time derivative of the heat kernel is bounded by a Gaussian-integrable polynomial on compact sets and positive time intervals; if these fail on a complete manifold with only $\mathrm{Ric}\ge 0$, the pointwise source equations and the propagation argument lose their foundation.

Editorial extensions

If this is right

The uniform estimate holds for every center and every radius, not merely for a fixed basepoint at infinity or for an $R$-dependent optimal center.
The exponent $n-2$ cannot be improved, since the sharp model $S^2_a\times\mathbb{R}^{n-2}$ attains the growth and satisfies the same curvature conditions.
Any complete noncompact manifold with $\mathrm{Ric}\ge 0$ and $\mathrm{Scal}\ge 1$ is forced to have at most $(n-2)$-dimensional macroscopic volume growth at all scales.
The argument identifies a concrete heat-flow quantity, the trace defect $E$, that carries the missing two dimensions; this gives a new analytic handle for volume growth questions under positive scalar curvature.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

One testable extension is whether the same two-unit source mechanism survives under weaker curvature bounds, such as $\mathrm{Ric}\ge -K$ or integral Ricci lower bounds; nothing in the structure forces the argument to be specific to $\mathrm{Ric}\ge 0$.
The proof suggests a local, scale-by-scale interpretation: regions with small $Q$ are 'Euclidean-like' in at most $n-2$ directions, and one could try to convert the heat-averaged source bound into Sobolev or isoperimetric inequalities rather than only volume bounds.
A numerical check on the sharp model should show $\lambda_a(t)\to 0$ on the sphere factor and $\Box_x E(t)\to 1$ as $t\to\infty$; verifying these asymptotics in a product example would be a cheap consistency test of the mechanism.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Read of arXiv:2608.13553. Take: if the proof is right—and I did not find a place where it isn't—this settles Gromov's 1986 question in the uniform form for all n≥3. That is a big deal. The new content is exactly the uniformity over centers and scales: WXZZ had the same exponent only for an infimum over centers, and Munteanu-Wang was n=3. Ge's Theorem 1.1 gives Vol(B(p,R)) ≤ C_n R^{n-2} for every p, every R, with the sharp model S^2×R^{n-2}. The mechanisms—heat covariance G, trace defect E, centered source Q, and the two-step propagation through □E=Q/2 and □(−S)=E/t^2—are clearly original and carefully set up. The paper is accurate that the pole-variable identities were formally present in Ni and Bamler, but not with this analytic weight.

What the paper does well. The algebraic first unit in Lemma 3.3 is just tr Ric ≥ 1 and 0≤G≤g; the second unit via weighted Weitzenböck on the Q<2−ε region is the real geometric input, and the spectral cutoff is handled cleanly. Proposition 5.2's entropy decay is short and the Tonelli swap is legitimate. Appendix A does a service by distinguishing pole-variable from output-variable identities. I line-checked the variance decomposition, Lemma 3.4, and the constants in Theorem 3.1; they cohere.

Soft spots. The load-bearing premise is Lemma B.1, uniform pole-variable heat-kernel jet bounds. Everything that lets pole derivatives pass under the output integrals leans on it, and it leans on Kotschwar's gradient estimate, Li-Yau Harnack, and terminal-time local Schauder estimates. I checked the key cancellation R(y) r_y = r_0 and the polynomial-in-y growth; it looks sound, and the constants are local as Remark B.2 claims. But this is exactly where I would want a second expert's audit before staking anything on it. The step from Lemma 3.9 to (3.19) also uses t≥6/ε and ε^{-2} assumptions that are coherent but should be re-read for edge cases. None of this amounts to a found error; the risk is un-audited dependence on noncompact regularity estimates.

Who it's for. Any geometer working on scalar curvature, heat kernels, or Gromov-type volume bounds. Yes, it deserves a serious referee—ideally one who knows Kotschwar's estimate and Schauder theory on complete manifolds. I would bring it to a reading group and would cite it if the domain overlaps my work.

Recommendation: send to peer review; the referee should be asked to scrutinize Lemma B.1 and the weighted Weitzenböck step in §3.4 before trusting the main theorem.

Referee Report

0 major / 5 minor

Summary. The paper proves Theorem 1.1: every complete, connected, noncompact n-dimensional Riemannian manifold with Ric_g >= 0 and Scal_g >= 1 satisfies Vol_g(B(p,R)) <= C_n R^{n-2} for all p in M and all R > 0, with C_n depending only on n. This is the uniform form of Gromov's 1986 volume-growth question. The proof introduces the heat-kernel Fisher covariance G_x(t) = 2t integral d_x h tensor d_x h dmu, the trace defect E_x(t) = (t/2)(n - tr g G_x(t)), the centered Hessian variance D, and the source Q, then derives the pointwise identities square_x E = Q/2 and square_x(-S) = E/t^2. The two-unit heat-averaged lower bound for Q (Theorem 3.1) is propagated through the minimal heat-potential comparison to obtain S_x(T) <= -log T + C_n, and the Li-Yau Gaussian estimate converts this entropy decay into the volume bound. Appendix B supplies the uniform pole-variable heat-kernel jet bounds used to justify differentiations under the output integrals.

Significance. If the result is correct, it settles a prominent open problem in geometric analysis, giving the uniform dimension-loss volume bound conjectured by Gromov and improving substantially on the n=3 result of Munteanu-Wang and the infimum-over-centers result of Wang-Xie-Zhu-Zhu. The proof is self-contained in the sense that the new objects G, E, D, and Q are defined directly from the heat kernel, no parameters are fitted, and the sharp model S^2_a x R^{n-2} is identified. I found no circular use of the target inequality and no algebraic error in the core chain from Proposition 2.9 through Lemma 3.3 and Theorem 3.1 to Proposition 5.2. The most delicate point, Lemma B.1 on uniform pole-variable heat-kernel jets, is load-bearing for the pointwise identities used throughout, and on line-level inspection the proof closes: the Kotschwar gradient estimate bounds the logarithmic ratio independently of the output point, and the local Schauder bootstrap contributes only polynomial factors absorbed by the Gaussian tail from Li-Yau.

minor comments (5)

[Proof of Theorem 1.1, Section 5] The displayed inequality "log Vol(B(x,sqrt(t)))/t^{n/2} <= S_x(t)+C_n" is dimensionally incorrect as written; it should read log(Vol(B(x,sqrt(t)))/t^{n/2}) <= S_x(t)+C_n, equivalently log Vol(B(x,sqrt(t))) - (n/2) log t <= S_x(t)+C_n. The subsequent conclusion for the volume bound is correct once this typo is repaired.
[Proposition 3.5] In the estimate for |B_t(p)|^2, the middle term should be (3/(2t))|g - G|^2 rather than (3/(2t))|G|^2, since it arises from |G/(2t) - g/(2t)|^2. The final bound |B_t(p)|^2 <= 6t D_p(t) + C_n/t is unaffected because G lies between 0 and g, but the intermediate line should be corrected.
[Lemma 3.8] The constant (n+1) in (3.14) appears to be a minor off-by-one from the crude bounds |d alpha|^2 <= 2|nabla alpha|^2 and |delta alpha|^2 <= n|nabla alpha|^2, which give (n+2). Since only a dimension constant C_n is needed, this does not affect the argument, but it should be adjusted for accuracy.
[Lemma B.1 and Remark B.2] The uniformity of the terminal-time interior Schauder estimate (B.6) over all x in K and all r in (0,r_0] is stated in one sentence. Because Lemma B.1 is the analytic foundation for Proposition 2.8 and hence for the pole-variable identities, the authors should add a few lines explaining the scaling argument in a fixed coordinate chart: the metric coefficients are uniformly smooth on a relatively compact neighborhood of K, so after rescaling a standard interior parabolic estimate has constants independent of x and r.
[Proposition 2.4] The limiting argument used to pass from the log-Sobolev inequality (2.7) to the exponential score bound (2.9) is sketched rather tersely. A brief comment on the dominated convergence used when chi_R -> 1 and epsilon -> 0 would improve readability, especially because this is the step that yields the fourth-moment bound used later in Proposition 3.5.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the proof is self-contained, uses no fitted parameters, and does not rely on self-citations.

full rationale

I walked the derivation chain from the definitions of G, E, D, and Q, through the pointwise identities □_x E = Q/2 and □_x(-S) = E/t^2, the heat-averaged two-unit source estimate Theorem 3.1, the entropy decay Proposition 5.2, and the final Li-Yau conversion to the volume bound. At each load-bearing step the target inequality does not appear as an input. The quantities G, E, D, and Q are defined directly from the heat kernel and the ambient metric, not from the volume-growth conclusion or from each other in a circular way: E is the trace defect of G, D is a centered Hessian variance, and Q is then shown by calculation in Proposition 2.9 to satisfy □_x E = Q/2. The two units in Theorem 3.1 come from the algebraic inequality in Lemma 3.3 under Scal ≥ 1 and 0 ≤ G ≤ g, plus the weighted Weitzenbock estimate under Ric ≥ 0; neither step assumes the volume bound. Proposition 5.2 is a pure propagation of the source lower bound through the minimal heat-potential comparison Lemma 5.1 and the semigroup property. The regularity Lemma B.1, while analytically the most delicate premise, is proved from external and standard ingredients—Kotschwar's gradient estimate, Li-Yau Harnack and Gaussian bounds, and interior parabolic Schauder estimates—and its failure would be a correctness defect, not a circularity. The paper contains no fitting of parameters to data, no renamed empirical pattern, and no uniqueness theorem imported from the author's own prior work; in fact the reference list contains no self-citations by Jian Ge. I therefore find no step in which an equation is equivalent to its input by construction, and no load-bearing self-citation chain.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No parameters are fitted: the proof is analytic and the dimensional constants C_n are existential constants derived from the estimates, not tuned to data. The auxiliary quantities G, E, D, and Q are new but are explicitly constructed from the heat kernel; they are not unexplained postulates. The relevant axioms are the curvature hypotheses and the standard noncompact heat kernel theorems invoked with precise citations.

assumptions (5)

domain assumption Complete, connected, noncompact (M,g) with Ric>=0 and Scal>=1, n>=3.
These are the hypotheses of Theorem 1.1; every step in Sections 2-5 operates under them.
standard math The minimal heat kernel H exists, is unique, positive, symmetric, smooth, and stochastically complete under Ric>=0.
Invoked in Section 2.1; ensures H(x,y,t)dVol is a probability measure and that the heat semigroup acts by integration.
standard math Bakry-Emery CD(0,infinity) reverse Poincare and logarithmic Sobolev inequalities from BGL14 hold under Ric>=0.
These yield the covariance bound 0<=G<=g and the exponential score bound (2.9), which are foundations for the trace defect and spectral gap.
standard math Li-Yau Harnack and Gaussian bounds, Kotschwar's pole-gradient estimate, the integrated Fisher bound, and local Schauder estimates hold on complete noncompact Ric>=0 manifolds.
Used in Lemma 4.1, Lemma 3.11, and Appendix B to control tails, justify differentiation under integrals, and supply dimension-only constants in Theorem 3.1.
standard math The minimal heat potential comparison principle for the inhomogeneous heat equation on an exhaustion, Lemma 5.1, and the existence of Greene-Wu proper exhaustion functions, Lemma 3.9.
Needed to propagate the two-unit source estimate from Q to E to -S without assuming boundary conditions at infinity.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Heat kernel geometry and Gromov's volume growth conjecture." pith.science (2026). https://pith.science/paper/ETO37DRX

@misc{pith2026260813553,
  author       = {Pith},
  title        = {Pith review of: Heat kernel geometry and Gromov's volume growth conjecture},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ETO37DRX}},
  note         = {Machine review of arXiv:2608.13553}
}

abstract

In 1986, Gromov asked whether every complete noncompact $n$-dimensional Riemannian manifold with nonnegative Ricci curvature and scalar curvature at least one satisfies: \[ \Vol_g (B(p, R))\le C_{n}R^{n-2} \] for all $p\in M$ and $R>0$. We answer this question affirmatively using the heat-kernel Fisher metric and Nash entropy.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

7 extracted references · 4 canonical work pages

[2]

New monotonicity formulas for Ricci curvature and applications; I

arXiv:1111.4715 [math.DG]. [Gri09] Alexander Grigor’yan.Heat Kernel and Analysis on Manifolds. Vol

work page Pith review arXiv
[12]

Provi- dence, RI: American Mathematical Society, 1996.doi:10.1090/ gsm/012

Graduate Studies in Mathematics. Provi- dence, RI: American Mathematical Society, 1996.doi:10.1090/ gsm/012. [Lie96] Gary M. Lieberman.Second Order Parabolic Differential Equa- tions. River Edge, NJ: World Scientific, 1996.doi:10.1142/

work page 1996
[47]

Large Riemannian manifolds

AMS/IP Studies in Advanced Mathematics. Providence, RI: American Mathematical Society and International Press, 2009.doi:10.1090/amsip/047. [Gro86] Mikhail Gromov. “Large Riemannian manifolds”. In:Curva- ture and Topology of Riemannian Manifolds (Katata, 1985). Ed. by Katsuhiro Shiohama, Takashi Sakai, and Toshikazu Sunada. Vol

doi:10.1090/amsip/047 2009
[348]

New monotonicity formulas for Ricci curvature and applications. I

Grundlehren der mathematischen Wissenschaften. Cham: Springer, 2014.doi:10.1007/978-3-319-00227-9. [Col12] Tobias Holck Colding. “New monotonicity formulas for Ricci curvature and applications. I”. In:Acta Mathematica209.2 (2012), pp. 229–263.doi:10 . 1007 / s11511 - 012 - 0086 -

doi:10.1007/978-3-319-00227-9 2012
[1201]

Hamilton's gradient estimate for the heat kernel on complete manifolds

Lecture Notes in Mathematics. Berlin: Springer, 1986, pp. 108–121.doi:10.1007/BFb0075649. [GW79] Robert E. Greene and Hung-Hsi Wu. “C ∞ Approximations of Convex, Subharmonic, and Plurisubharmonic Functions”. In: Annales scientifiques de l’ ´Ecole Normale Sup´ erieure. 4th ser. 12.1 (1979), pp. 47–84.doi:10.24033/asens.1361. [ISS08] Mitsuhiro Itoh, Hiroyas...

work page Pith review arXiv 1979
[2020]

[BGL14] Dominique Bakry, Ivan Gentil, and Michel Ledoux.Analy- sis and Geometry of Markov Diffusion Operators

arXiv:2008.07093 [math.DG]. [BGL14] Dominique Bakry, Ivan Gentil, and Michel Ledoux.Analy- sis and Geometry of Markov Diffusion Operators. Vol

arXiv 2008
[3302]

The entropy formula for linear heat equation

[LY86] Peter Li and Shing-Tung Yau. “On the parabolic kernel of the Schr¨ odinger operator”. In:Acta Mathematica156.3–4 (1986), pp. 153–201.doi:10.1007/BF02399203. REFERENCES 33 [MW26] Ovidiu Munteanu and Jiaping Wang. “Geometry of three- dimensional manifolds with positive scalar curvature”. In: American Journal of Mathematics148.1 (2026), pp. 131– 160.d...

work page Pith review arXiv 1986

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [2]

New monotonicity formulas for Ricci curvature and applications; I

arXiv:1111.4715 [math.DG]. [Gri09] Alexander Grigor’yan.Heat Kernel and Analysis on Manifolds. Vol

work page Pith review arXiv

[2] [12]

Provi- dence, RI: American Mathematical Society, 1996.doi:10.1090/ gsm/012

Graduate Studies in Mathematics. Provi- dence, RI: American Mathematical Society, 1996.doi:10.1090/ gsm/012. [Lie96] Gary M. Lieberman.Second Order Parabolic Differential Equa- tions. River Edge, NJ: World Scientific, 1996.doi:10.1142/

work page 1996

[3] [47]

Large Riemannian manifolds

AMS/IP Studies in Advanced Mathematics. Providence, RI: American Mathematical Society and International Press, 2009.doi:10.1090/amsip/047. [Gro86] Mikhail Gromov. “Large Riemannian manifolds”. In:Curva- ture and Topology of Riemannian Manifolds (Katata, 1985). Ed. by Katsuhiro Shiohama, Takashi Sakai, and Toshikazu Sunada. Vol

doi:10.1090/amsip/047 2009

[4] [348]

New monotonicity formulas for Ricci curvature and applications. I

Grundlehren der mathematischen Wissenschaften. Cham: Springer, 2014.doi:10.1007/978-3-319-00227-9. [Col12] Tobias Holck Colding. “New monotonicity formulas for Ricci curvature and applications. I”. In:Acta Mathematica209.2 (2012), pp. 229–263.doi:10 . 1007 / s11511 - 012 - 0086 -

doi:10.1007/978-3-319-00227-9 2012

[5] [1201]

Hamilton's gradient estimate for the heat kernel on complete manifolds

Lecture Notes in Mathematics. Berlin: Springer, 1986, pp. 108–121.doi:10.1007/BFb0075649. [GW79] Robert E. Greene and Hung-Hsi Wu. “C ∞ Approximations of Convex, Subharmonic, and Plurisubharmonic Functions”. In: Annales scientifiques de l’ ´Ecole Normale Sup´ erieure. 4th ser. 12.1 (1979), pp. 47–84.doi:10.24033/asens.1361. [ISS08] Mitsuhiro Itoh, Hiroyas...

work page Pith review arXiv 1979

[6] [2020]

[BGL14] Dominique Bakry, Ivan Gentil, and Michel Ledoux.Analy- sis and Geometry of Markov Diffusion Operators

arXiv:2008.07093 [math.DG]. [BGL14] Dominique Bakry, Ivan Gentil, and Michel Ledoux.Analy- sis and Geometry of Markov Diffusion Operators. Vol

arXiv 2008

[7] [3302]

The entropy formula for linear heat equation

[LY86] Peter Li and Shing-Tung Yau. “On the parabolic kernel of the Schr¨ odinger operator”. In:Acta Mathematica156.3–4 (1986), pp. 153–201.doi:10.1007/BF02399203. REFERENCES 33 [MW26] Ovidiu Munteanu and Jiaping Wang. “Geometry of three- dimensional manifolds with positive scalar curvature”. In: American Journal of Mathematics148.1 (2026), pp. 131– 160.d...

work page Pith review arXiv 1986