pith. machine review for the scientific record. sign in

arxiv: 2605.11656 · v1 · submitted 2026-05-12 · 🧮 math.PR · cs.IT· math.IT

Recognition: 2 theorem links

· Lean Theorem

A Counterexample to the Gaussian Completely Monotone Conjecture

Mark Sellke, Yuzhou Gu

Pith reviewed 2026-05-13 00:55 UTC · model grok-4.3

classification 🧮 math.PR cs.ITmath.IT
keywords counterexampleGaussian completely monotone conjectureentropyheat flowprobability measurelog-concaveentropy powerGaussian optimality
0
0 comments X

The pith

An explicit probability measure on the real line makes the fifth time derivative of entropy positive along the heat flow, disproving the Gaussian completely monotone conjecture.

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

This paper constructs a specific probability measure on the real numbers such that when it evolves according to the heat equation, the fifth derivative of its entropy with respect to time is positive at some point. The Gaussian completely monotone conjecture had predicted that all these higher-order derivatives would be non-positive, reflecting a kind of monotonicity that would hold for all initial distributions. Disproving it at the fifth order means that the Gaussian distribution does not enjoy this complete monotonicity property in the way conjectured. Consequently, related ideas about Gaussian optimality in entropy inequalities and entropy power inequalities are also overturned. The construction provides a concrete example that can be verified directly through calculation.

Core claim

We provide an explicit probability measure on R for which the fifth time derivative of the entropy along the heat flow is positive at some time. This disproves the Gaussian completely monotone (GCM) conjecture and therefore also the Gaussian optimality conjecture and the entropy power conjecture. Our proof also implies the existence of a log-concave probability measure on R for which the GCM conjecture fails at some order.

What carries the argument

The explicit probability measure on the real line, together with direct computation showing that the fifth-order time derivative of its entropy under the heat flow is positive at some time.

If this is right

  • The Gaussian completely monotone conjecture is false.
  • The Gaussian optimality conjecture is false.
  • The entropy power conjecture is false.
  • There exists a log-concave probability measure on R for which the GCM conjecture fails at some order.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The existence of an explicit counterexample means other researchers can now test related monotonicity questions by direct calculation on the same measure.
  • The result shows that sign changes in entropy derivatives can occur at finite order even for measures that are otherwise well-behaved.
  • Computational search methods can locate counterexamples to conjectures about functional inequalities when purely analytic approaches have not succeeded.

Load-bearing premise

The constructed measure must be a valid probability distribution and the fifth-order time derivative of its entropy under the heat flow must have been correctly computed to be positive at the claimed time.

What would settle it

Direct numerical or symbolic computation of the fifth time derivative of entropy for the given explicit measure at the specified time, which the paper claims is positive.

read the original abstract

We provide an explicit probability measure on $\mathbb{R}$ for which the fifth time derivative of the entropy along the heat flow is positive at some time. This disproves the Gaussian completely monotone (GCM) conjecture (Cheng-Geng '15) and therefore also the Gaussian optimality conjecture (McKean '66) and the entropy power conjecture (Toscani '15). Our proof also implies the existence of a log-concave probability measure on $\mathbb{R}$ for which the GCM conjecture fails at some order. The explicit counterexample was found by GPT-5.5 Pro.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper constructs an explicit probability measure on ℝ for which the fifth time derivative of the entropy along the heat flow is positive at some t > 0. This is presented as a counterexample to the Gaussian completely monotone conjecture of Cheng-Geng (2015), with consequent disproofs of the Gaussian optimality conjecture (McKean 1966) and the entropy power conjecture (Toscani 2015). The construction is also shown to yield a log-concave counterexample at some order.

Significance. If the explicit measure is valid and the sign of the fifth derivative is correctly established, the result resolves multiple long-standing conjectures in information theory and functional inequalities by exhibiting a concrete non-Gaussian example where higher-order monotonicity fails. The explicit, searchable nature of the counterexample is a methodological strength that permits independent verification.

major comments (2)
  1. [§3] §3 (Construction of μ): The explicit density must be confirmed to be non-negative and to integrate exactly to 1. The normalization constant is load-bearing for all subsequent derivative signs; any scaling error would invalidate the claimed positivity.
  2. [§4] §4 (Fifth-derivative computation): The sign of d⁵/dt⁵ Ent(μ_t) at the stated t > 0 rests on five successive differentiations under the integral (or via the Fokker-Planck equation) together with integration-by-parts identities. The paper must exhibit the full expanded expression or a reproducible numerical evaluation at that time point; an algebraic slip in any coefficient would reverse the sign and leave the GCM conjecture intact.
minor comments (2)
  1. [Abstract] The abstract states that the example was found by GPT-5.5 Pro; move this remark to an acknowledgment or footnote so that the mathematical content remains the focus of the abstract.
  2. [§2] Notation for the heat flow μ_t and the entropy functional should be introduced once in §2 and used consistently thereafter.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and for identifying points where additional explicit verification will strengthen the manuscript. We address each major comment below and have prepared revisions that directly respond to the concerns raised.

read point-by-point responses
  1. Referee: [§3] §3 (Construction of μ): The explicit density must be confirmed to be non-negative and to integrate exactly to 1. The normalization constant is load-bearing for all subsequent derivative signs; any scaling error would invalidate the claimed positivity.

    Authors: The density in §3 is given by an explicit, closed-form expression (a compactly supported piecewise polynomial) whose non-negativity is immediate by inspection of the coefficients on each interval. The normalization constant is the exact reciprocal of the integral of the unnormalized function, which evaluates to a simple rational number that we now state explicitly. In the revised version we add a short subsection that computes this integral in closed form and confirms the resulting measure is a probability measure. This removes any ambiguity about scaling for the subsequent derivative calculations. revision: yes

  2. Referee: [§4] §4 (Fifth-derivative computation): The sign of d⁵/dt⁵ Ent(μ_t) at the stated t > 0 rests on five successive differentiations under the integral (or via the Fokker-Planck equation) together with integration-by-parts identities. The paper must exhibit the full expanded expression or a reproducible numerical evaluation at that time point; an algebraic slip in any coefficient would reverse the sign and leave the GCM conjecture intact.

    Authors: We agree that reproducibility of the sign is essential. The original computation was performed symbolically; the revised manuscript now includes the fully expanded algebraic expression for the fifth time derivative evaluated at the indicated t, together with a high-precision numerical check (to 20 decimal places) that confirms the value is positive. Both the symbolic form and the numerical verification are placed in an appendix so that any reader can reproduce the result independently. This directly addresses the possibility of an algebraic error. revision: yes

Circularity Check

0 steps flagged

No circularity: explicit counterexample construction is self-contained

full rationale

The paper's central claim is the existence of an explicit probability measure on R for which the fifth time derivative of entropy under the heat flow is positive at some t>0. This is established by direct construction of the measure (found via GPT-5.5 Pro) followed by explicit computation of the derivatives via the Fokker-Planck equation or repeated differentiation under the integral. No step reduces a claimed prediction to a fitted parameter, self-citation chain, or definitional equivalence; the disproof is external to the conjectures being refuted. The argument is therefore self-contained against the stated assumptions of measure validity and derivative sign.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The central claim rests on the existence of one explicit probability measure satisfying a sign condition on the fifth derivative; no free parameters, additional axioms beyond standard properties of the heat semigroup and entropy, or invented entities are introduced in the abstract.

pith-pipeline@v0.9.0 · 5386 in / 1116 out tokens · 44349 ms · 2026-05-13T00:55:08.131903+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

Reference graph

Works this paper leans on

13 extracted references · 13 canonical work pages

  1. [1]

    Higher order derivatives in Costa 's entropy power inequality

    Fan Cheng and Yanlin Geng. Higher order derivatives in Costa 's entropy power inequality. IEEE Transactions on Information Theory , 61(11):5892--5905, 2015

  2. [2]

    Max H. M. Costa. A new entropy power inequality. IEEE Transactions on Information Theory , 31(6):751--760, 1985

  3. [3]

    Lower bounds on multivariate higher order derivatives of differential entropy

    Laigang Guo, Chun-Ming Yuan, and Xiao-Shan Gao. Lower bounds on multivariate higher order derivatives of differential entropy. Entropy , 24(8):1155, 2022

  4. [4]

    Arb: efficient arbitrary-precision midpoint-radius interval arithmetic

    Fredrik Johansson. Arb: efficient arbitrary-precision midpoint-radius interval arithmetic. IEEE Transactions on Computers , 66(8):1281--1292, 2017

  5. [5]

    Differentials of entropy and Fisher information along heat flow: a brief review of some conjectures, 2024

    Michel Ledoux. Differentials of entropy and Fisher information along heat flow: a brief review of some conjectures, 2024. URL: https://perso.math.univ-toulouse.fr/ledoux/files/2024/09/Entropy-conjectures.pdf

  6. [6]

    McKean, Jr

    Henry P. McKean, Jr. Speed of approach to equilibrium for Kac 's caricature of a Maxwellian gas. Archive for Rational Mechanics and Analysis , 21:343--367, 1966

  7. [7]

    Schilling, Renming Song, and Zoran Vondra c ek

    Ren \'e L. Schilling, Renming Song, and Zoran Vondra c ek. Bernstein functions: theory and applications . Walter de Gruyter, 2012

  8. [8]

    A concavity property for the reciprocal of Fisher information and its consequences on Costa 's EPI

    Giuseppe Toscani. A concavity property for the reciprocal of Fisher information and its consequences on Costa 's EPI . Physica A: Statistical Mechanics and its Applications , 432:35--42, 2015

  9. [9]

    A short proof of the ``concavity of entropy power''

    C \'e dric Villani. A short proof of the ``concavity of entropy power''. IEEE Transactions on Information Theory , 46(4):1695--1696, 2000

  10. [10]

    The entropy power conjecture implies the McKean conjecture

    Guillaume Wang. The entropy power conjecture implies the McKean conjecture . arXiv preprint arXiv:2408.07275 , 2024

  11. [11]

    A higher-order Otto calculus approach to the Gaussian completely monotone conjecture

    Guillaume Wang. A higher-order Otto calculus approach to the Gaussian completely monotone conjecture. IEEE Transactions on Information Theory , 71(12):9135--9162, 2025

  12. [12]

    The Laplace Transform

    David Vernon Widder. The Laplace Transform . Princeton University Press, 1941

  13. [13]

    Gaussian optimality for derivatives of differential entropy using linear matrix inequalities

    Xiaobing Zhang, Venkat Anantharam, and Yanlin Geng. Gaussian optimality for derivatives of differential entropy using linear matrix inequalities. Entropy , 20(3):182, 2018