arxiv: 2605.08240 · v1 · submitted 2026-05-07 · 🧮 math.DG

Recognition: 2 theorem links

· Lean Theorem

Geometry of tangent bundles of statistical manifolds equiped with Cheeger-Gromoll type metrics

Esmaeil Peyghan, Leila Nourmohammadifar

Pith reviewed 2026-05-12 01:10 UTC · model grok-4.3

classification 🧮 math.DG

keywords statistical manifoldstangent bundleCheeger-Gromoll metricLevi-Civita connectionsectional curvatureskewness tensorgeodesicsinformation geometry

0 comments

The pith

The tangent bundle of a statistical manifold with generalized Cheeger-Gromoll metrics has curvature determined by the base Riemannian curvature and skewness tensor, with explicit conditions for constant sectional curvature.

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

This paper investigates the tangent bundle TM over a statistical manifold equipped with a two-parameter family of generalized Cheeger-Gromoll metrics. It derives the Levi-Civita connection and expresses the curvature tensor using the base manifold's curvature together with the skewness tensor. The analysis covers geodesics, totally geodesic fibers, incompressible flows, and gives necessary and sufficient conditions under which TM itself has constant sectional curvature. A sympathetic reader would care because these constructions lift statistical geometry to the bundle level, allowing explicit curvature calculations in models such as normal distributions.

Core claim

The curvature of (TM, g_{p,q}) is expressed in terms of the Riemannian curvature and skewness tensor K of the base statistical manifold (M, g, ∇). Necessary and sufficient conditions are established for TM to admit constant sectional curvature, with explicit formulas for sectional curvatures in horizontal, vertical, and mixed directions that also yield the scalar curvature.

What carries the argument

The two-parameter family of generalized Cheeger-Gromoll metrics g_{p,q} on TM, which induces a decomposition into horizontal and vertical subspaces and permits curvature formulas involving the base skewness tensor K.

If this is right

The fibers of TM are totally geodesic for appropriate choices of p and q.
The geodesic flow on TM is incompressible under specific parameter restrictions.
Geodesics on TM can be lifted from base geodesics using the computed connection.
Scalar curvature of TM follows from summing the explicit sectional curvature expressions in all planes.

Where Pith is reading between the lines

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

The constant-curvature cases may correspond to rigid statistical manifolds such as certain exponential families.
These bundle metrics could define new divergence functions that incorporate first-order information from tangent vectors.
The framework suggests similar curvature computations for other bundle constructions in information geometry beyond the tangent bundle.

Load-bearing premise

That the two-parameter family g_{p,q} defines a smooth positive-definite Riemannian metric on TM for the given range of parameters.

What would settle it

Direct computation of the sectional curvature on the tangent bundle of the normal distributions manifold for concrete p and q values, checking whether it is constant precisely when the stated conditions on the base manifold hold.

read the original abstract

In this paper, we investigate the geometry of the tangent bundle $TM$ of a statistical manifold $(M,g,\nabla)$ endowed with a two-parameter family of generalized Cheeger--Gromoll metrics $g_{p,q}$. We compute the associated the Levi--Civita connection $\nabla^{p,q}$ and express its curvature in terms of the Riemannian curvature and the skewness tensor $K$ of the base statistical manifold. We further analyze the behavior of geodesics, identify conditions under which the fibers of $TM$ are totally geodesic, and determine when the geodesic flow associated with $g_{p,q}$ is incompressible. Moreover, we establish necessary and sufficient conditions for the tangent bundle to admit constant sectional curvature. Several examples are provided to illustrate the theory, including statistically deformed Euclidean spaces and information geometric models such as the manifold of normal distributions. The sectional curvature of $(TM, g_{p,q})$ is computed for horizontal, vertical, and mixed directions, leading to a concise expression for the corresponding scalar curvature.

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.

Desk Editor's Note private letter to a colleague

The paper derives explicit curvature formulas and constant sectional curvature conditions for tangent bundles over statistical manifolds using a two-parameter Cheeger-Gromoll metric family, but positive definiteness of the metric needs direct verification when the skewness tensor is nonzero.

read the letter

The main contribution is the curvature decomposition on TM that brings in the skewness tensor K from the statistical structure, along with necessary and sufficient conditions for constant sectional curvature and some statements about geodesics and incompressible flow. They also compute the Levi-Civita connection for the two-parameter metric g_{p,q} and give sectional curvatures in the horizontal, vertical, and mixed directions, plus scalar curvature. The examples on deformed Euclidean space and the normal distributions manifold make the formulas testable rather than purely formal. That is the concrete work the paper actually delivers. The derivations follow the usual coordinate or frame computations that appear in tangent bundle papers, so the algebra itself looks consistent with standard Riemannian geometry techniques. The soft spot is the claim that g_{p,q} is Riemannian for the stated range of p and q. When K is nonzero the horizontal-vertical splitting picks up cross terms from the difference between the statistical connection and the Levi-Civita connection. The abstract asserts positivity but does not display an explicit matrix representation or eigenvalue bound that accounts for those terms. If the full paper contains a direct check of the quadratic form, the later results stand; otherwise the curvature expressions and constant-curvature conditions only hold reliably in the K=0 case. This is a specialized paper for readers already comfortable with both statistical manifolds and tangent bundle geometry. Someone working in information geometry who wants explicit curvature expressions on TM will get usable formulas and examples. It is worth sending to a referee who can confirm the metric positivity step and check the geodesic conditions against the provided examples.

Referee Report

1 major / 2 minor

Summary. The manuscript investigates the geometry of the tangent bundle TM of a statistical manifold (M, g, ∇) equipped with a two-parameter family of generalized Cheeger-Gromoll metrics g_{p,q}. It derives the Levi-Civita connection ∇^{p,q}, expresses the curvature tensor in terms of the base Riemannian curvature and the skewness tensor K, analyzes geodesics (including conditions for totally geodesic fibers and incompressible geodesic flow), and establishes necessary and sufficient conditions for (TM, g_{p,q}) to have constant sectional curvature. Explicit computations of sectional curvatures in horizontal, vertical, and mixed directions are given, along with the scalar curvature; several examples, including statistically deformed Euclidean spaces and the manifold of normal distributions, are provided to illustrate the results.

Significance. If the metric g_{p,q} is indeed Riemannian and the derivations hold, the work extends classical results on tangent bundle geometries to the setting of statistical manifolds, which are fundamental in information geometry. The explicit curvature formulas in terms of base curvature and K, together with the characterization of constant sectional curvature, could serve as tools for studying geometric properties of information manifolds and related statistical models. The provision of concrete examples adds applicability.

major comments (1)

[Definition of g_{p,q}] Definition of the metric g_{p,q} (presumably §2 or §3): The claim that g_{p,q} is a smooth Riemannian metric on TM for the given range of p and q is load-bearing for the entire paper, as all subsequent results on ∇^{p,q}, curvature, geodesics, and constant sectional curvature presuppose a well-defined Riemannian structure. When the skewness tensor K is nonzero, the horizontal-vertical splitting introduces cross terms in the quadratic form defining g_{p,q}; the manuscript supplies neither an explicit matrix representation in adapted frames nor an eigenvalue estimate that accounts for these terms to confirm positive-definiteness.

minor comments (2)

[Introduction] The abstract and introduction could more explicitly reference prior work on Cheeger-Gromoll metrics and statistical connections to clarify the novelty of the two-parameter deformation.
[Examples] In the examples section, ensure that the skewness tensor K is computed explicitly for each model (e.g., normal distributions) so that readers can verify the curvature expressions.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful reading of our manuscript and the constructive comment on the positive-definiteness of the metric. We address the point below and will revise the paper accordingly to strengthen the foundational claims.

read point-by-point responses

Referee: Definition of g_{p,q} (presumably §2 or §3): The claim that g_{p,q} is a smooth Riemannian metric on TM for the given range of p and q is load-bearing for the entire paper, as all subsequent results on ∇^{p,q}, curvature, geodesics, and constant sectional curvature presuppose a well-defined Riemannian structure. When the skewness tensor K is nonzero, the horizontal-vertical splitting introduces cross terms in the quadratic form defining g_{p,q}; the manuscript supplies neither an explicit matrix representation in adapted frames nor an eigenvalue estimate that accounts for these terms to confirm positive-definiteness.

Authors: We agree that a rigorous verification of positive-definiteness is essential, particularly when the skewness tensor K induces cross terms between the horizontal and vertical distributions. The metric g_{p,q} is defined in Section 2 via the natural lifts with respect to the affine connection ∇ of the statistical manifold, extending the classical Cheeger-Gromoll construction. In the revised manuscript we will add a new lemma immediately following the definition, which: (i) gives the explicit matrix representation of g_{p,q} in an adapted local frame {e_i^h, e_i^v} (accounting for the cross terms proportional to K), (ii) computes the associated quadratic form, and (iii) proves positive-definiteness for the stated ranges of p and q by showing that the eigenvalues remain strictly positive; the perturbation arising from K is controlled using the tensorial symmetries of K and standard bounds on compact subsets of TM. This addition will be placed before any curvature or geodesic computations, ensuring the Riemannian structure is fully justified. revision: yes

Circularity Check

0 steps flagged

No circularity: curvature formulas are direct computations from the defined metric and base data

full rationale

The paper defines the two-parameter metric family g_{p,q} on TM, then applies the standard Levi-Civita connection formula and curvature tensor expression (via Koszul or coordinate methods) to obtain ∇^{p,q} and its sectional curvatures in terms of the base Riemannian curvature and the given skewness tensor K. These are algebraic identities following from the metric coefficients and the statistical connection; they do not reduce by construction to fitted parameters, self-definitions, or prior self-citations. The necessary-and-sufficient conditions for constant sectional curvature are obtained by setting the derived curvature components equal to a constant and solving, which is an independent algebraic step. No load-bearing self-citation chains or ansatz smuggling appear in the derivation. The work is a standard explicit computation in Riemannian geometry on bundles and remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The work rests on the standard definition of a statistical manifold and the assumption that g_{p,q} is a Riemannian metric; no free parameters, new entities, or ad-hoc axioms are introduced in the abstract.

axioms (2)

domain assumption The base space is a statistical manifold (M, g, ∇) equipped with a compatible skewness tensor K.
Invoked throughout the abstract as the starting structure on which the bundle metric is defined.
domain assumption The two-parameter family g_{p,q} defines a smooth Riemannian metric on the tangent bundle TM.
Required for the Levi-Civita connection and curvature computations to be well-defined.

pith-pipeline@v0.9.0 · 5483 in / 1481 out tokens · 70679 ms · 2026-05-12T01:10:16.377239+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

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

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear
We compute the associated Levi-Civita connection ∇^{p,q} and express its curvature in terms of the Riemannian curvature and the skewness tensor K of the base statistical manifold.
IndisputableMonolith/Foundation/AlexanderDuality.lean alexander_duality_circle_linking unclear
establish necessary and sufficient conditions for the tangent bundle to admit constant sectional curvature

Reference graph

Works this paper leans on

22 extracted references · 22 canonical work pages

[1]

Abbassi and M

M. Abbassi and M. Sarih,Killing Vector Fields on Tangent Bundles with Cheeger Gromoll Metric, Tsukuba J. Math., 27(2) (2003), 295-306

work page 2003
[2]

Amari,Information geometry of the EM and em algorithms for neural networks, Neural Networks8(9) (1995), 1379–1408

S. Amari,Information geometry of the EM and em algorithms for neural networks, Neural Networks8(9) (1995), 1379–1408

work page 1995
[3]

Amari,Natural Gradient Works Efficiently in Learning, Neural Computation,10(2) (1998), 251–276

S. Amari,Natural Gradient Works Efficiently in Learning, Neural Computation,10(2) (1998), 251–276

work page 1998
[4]

Amari and H

S. Amari and H. Nagaoka,Methods of information geometry, American Mathematical Society, 2000

work page 2000
[5]

Barbosu,Sasaki metric and the phase space, Carpathian Journal of Mathematics,14(1998), 139–142

M. Barbosu,Sasaki metric and the phase space, Carpathian Journal of Mathematics,14(1998), 139–142

work page 1998
[6]

L. P. Bettmann and J. Goold,Information geometry approach to quantum stochastic thermodynamics, Phys. Rev. E, 111(2025), 014133. DOI: https://doi.org/10.1103/PhysRevE.111.014133

work page doi:10.1103/physreve.111.014133 2025
[7]

Belkin, P

M. Belkin, P. Niyogi and V. Sindhwani,Manifold regularization: a geometric framework for learning from labeled and unlabeled examples, Journal of Machine Learning Research,7(2006), 2399-2434

work page 2006
[8]

Benyounes, E

M. Benyounes, E. Loubeau and C.M. Wood,Harmonic sections of Riemannian vector bundles, and metrics of Cheeger- Gromoll type, Diff. Geom. Appl.,25(2007), 322–334

work page 2007
[9]

Blaga, and C

A.M. Blaga, and C. Crasmareanu,Statistical Structures in Almost Paracontact Geometry, Bulletin of the Iranian Mathematical Society44(2018), 1407-1413

work page 2018
[10]

Calin and C

O. Calin and C. Udri¸ ste,Geometric Modeling in Probability and Statistics, Springer, Cham, Switzerland 2014

work page 2014
[11]

D. Cai, X. Liu, and L. Zhang,Inequalities on generalized normalizedδ-Casorati curvatures for submanifolds in statistical manifolds of constant curvatures, (Chinese) J. Jilin Univ. Sci.57(2) (2019), 206-212

work page 2019
[12]

Cheeger and D

J. Cheeger and D. Gromoll,On the structure of complete manifolds of nonnegative curvature, Ann. Math.96(1972), 413-443

work page 1972
[13]

Di Sipio,Rethinking LLM training through information geometry and quantum metrics, arXiv:2506.15830

R. Di Sipio,Rethinking LLM training through information geometry and quantum metrics, arXiv:2506.15830

work page arXiv
[14]

R. A. Fisher,On the mathematical foundations of theoretical statistics, Phil. Trans. Roy. Soc. London,222(1922), 309–368

work page 1922
[15]

Furuhata, I

H. Furuhata, I. Hasegawa, Y. Okuyama, K. Sato and M.H. Shahid,Sasakian statistical manifolds, J. Geom. Phys.117 (2017), 179-–186

work page 2017
[16]

Opozda,On the tangent bundles of statistical manifolds, In Proceedings of the Geometric Science of Information GSI 2023, St

B. Opozda,On the tangent bundles of statistical manifolds, In Proceedings of the Geometric Science of Information GSI 2023, St. Malo, France, 30 August–1 September 2023

work page 2023
[17]

Peyghan, L

E. Peyghan, L. Nourmohammadifar and D. Iosifidis,Statistical conformal Killing vector fields for FLRW spacetime, Phys Scr.99, 065253, (2024)

work page 2024
[18]

Peyghan, L

E. Peyghan, L. Nourmohammadifar and S. Uddin,Musical Isomorphisms and Statistical Manifolds, Mediterr. J. Math. (2022), Doi.org/10.1007/s00009-022-02141-z

work page doi:10.1007/s00009-022-02141-z 2022
[19]

Peyghan, L

E. Peyghan, L. Nourmohammadifar, and I. Mihai,Statistical Submanifolds Equipped withF-Statistical Connections, Mathematics,12(16) (2024), 2492. https://doi.org/10.3390/math12162492

work page doi:10.3390/math12162492 2024
[20]

Peyghan, F

E. Peyghan, F. L. Nourmohammadi, A. Tayebi,Cheeger–Gromoll type metrics on the (1,1)-tensor bundles, J. Contemp. Math. Anal.,48(6), (2013), 59-70

work page 2013
[21]

Opozda,Bochner’s technique for statistical structures, Ann

B. Opozda,Bochner’s technique for statistical structures, Ann. Global Anal. Geom.48(2015), 357–395

work page 2015
[22]

Sasaki,On the differential geometry of tangent bundles of Riemannian manifolds I, Tohoku Mathematical Journal, 10(3) (1958), 338–354

S. Sasaki,On the differential geometry of tangent bundles of Riemannian manifolds I, Tohoku Mathematical Journal, 10(3) (1958), 338–354. Department of Mathematics, F aculty of Science, Arak University, Arak, 38156-8-8349, Iran. Email address:e-peyghan@araku.ac.ir, l-nourmohammadifar@araku.ac.ir

work page 1958