arxiv: 2605.13004 · v1 · submitted 2026-05-13 · 🧮 math.PR · math.ST· stat.TH

Recognition: 2 theorem links

· Lean Theorem

Orientation in Poisson Cluster Processes via Imaginary Bispectra

Conor Kresin , Yifu Tang , Boris Baeumer , Ting Wang

Authors on Pith no claims yet

Pith reviewed 2026-05-14 20:34 UTC · model grok-4.3

classification 🧮 math.PR math.STstat.TH

keywords Poisson cluster processesfactorial bispectrumthird cumulantorientation detectionBartlett spectrumHawkes processespoint processesreversible null models

0 comments

The pith

A nonzero imaginary factorial bispectrum certifies orientation for stationary Poisson branching clusters when the reduced third cumulant is L1-integrable.

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

The paper examines what information survives when the direction of clustering is erased in one-sided Poisson cluster processes. It builds reversible null models that keep the same intensity and Bartlett spectrum, demonstrating that second-order statistics alone cannot recover temporal direction. For stationary Poisson branching clusters it derives the Fourier-Stieltjes transform of the reduced third cumulant and proves that, under an L1 integrability condition, a nonzero imaginary factorial bispectrum is enough to certify the original orientation. This matters for any directional point process where only the observed times are available and direction must be inferred rather than assumed.

Core claim

For stationary Poisson branching clusters in the L1 third-cumulant regime, the Fourier-Stieltjes transform of the reduced third cumulant yields a nonzero imaginary factorial bispectrum that certifies the direction of clustering. Matched reversible cluster nulls exist that preserve intensity and the full Bartlett spectrum, showing second-order structure is insufficient to identify orientation. Explicit orientation-erased nulls and reversible spectral matches are also given for monotone Hawkes kernels.

What carries the argument

The imaginary factorial bispectrum, obtained as the Fourier-Stieltjes transform of the reduced third cumulant of the cluster process.

If this is right

Second-order measures such as the Bartlett spectrum cannot distinguish temporal direction in these processes.
Reversible cluster nulls that match intensity and full spectrum can be constructed explicitly.
Monotone Hawkes kernels admit reversible spectral matches after orientation erasure.
Finite-window third-order contrasts provide practical orientation tests.

Where Pith is reading between the lines

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

The same bispectral signature might appear in other directed point processes once their third-order structure is examined.
Data analysts could test observed event sequences for hidden orientation by checking whether the imaginary bispectrum is nonzero.
The method supplies a concrete way to build surrogate data sets that preserve second-order statistics while removing direction.

Load-bearing premise

The process must be a stationary Poisson branching cluster whose reduced third cumulant satisfies the L1 integrability condition.

What would settle it

An explicit stationary Poisson branching cluster with L1 third cumulant whose imaginary factorial bispectrum is identically zero.

read the original abstract

We study what remains detectable about one-sided Poisson cluster processes after cluster orientation is erased. We construct matched reversible cluster nulls preserving intensity and the full Bartlett spectrum, showing that second-order structure alone need not identify temporal direction. For stationary Poisson branching clusters, we derive the Fourier--Stieltjes transform of the reduced third cumulant and show that, in the $L^1$ third-cumulant regime, a nonzero imaginary factorial bispectrum certifies orientation. We also give explicit orientation-erased nulls, reversible spectral matches for monotone Hawkes kernels, and finite-window third-order orientation contrasts.

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

Second-order spectra like the Bartlett one can be matched by reversible nulls, so the imaginary factorial bispectrum is needed to certify orientation in these cluster processes under the L1 regime.

read the letter

The main thing to know is that this paper constructs reversible Poisson cluster nulls that preserve both intensity and the full Bartlett spectrum, proving that second-order information alone cannot detect temporal direction. It then derives the Fourier-Stieltjes transform of the reduced third cumulant for stationary Poisson branching clusters and shows that a nonzero imaginary factorial bispectrum certifies orientation when the third cumulant satisfies the L1 condition. That construction and the third-order certificate are the actual new pieces. The nulls are explicit and the argument stays within the stated regime without hidden parameters or fitting steps. The paper also supplies orientation-erased nulls for monotone Hawkes kernels and some finite-window contrasts, which are practical touches. The L1 integrability requirement is a genuine restriction that will rule out heavier-tailed or longer-range cluster structures, but the paper states the condition plainly and does not claim more. Within the regime the derivation looks consistent and the reversible matching is a clean way to isolate what second-order methods miss. This is for readers already working with point-process spectra who need a direction diagnostic. It deserves peer review because the core construction is original, scoped correctly, and the math is grounded in the cumulant measures rather than simulation or approximation.

Referee Report

1 major / 2 minor

Summary. The paper studies detectability of orientation in one-sided Poisson cluster processes after orientation erasure. It constructs reversible cluster nulls that preserve intensity and the full Bartlett spectrum, showing second-order structure is insufficient to identify temporal direction. For stationary Poisson branching clusters in the L^1 third-cumulant regime, the Fourier-Stieltjes transform of the reduced third cumulant is derived, and a nonzero imaginary factorial bispectrum is shown to certify orientation. Explicit orientation-erased nulls, reversible spectral matches for monotone Hawkes kernels, and finite-window third-order orientation contrasts are also given.

Significance. If the central derivation holds, the result supplies a concrete higher-order certificate for temporal asymmetry in point processes where second-order spectra are matched by reversible nulls. The explicit construction of matched reversible nulls and the isolation of the imaginary bispectrum component are strengths that could inform applications in directed event modeling.

major comments (1)

The L^1 integrability condition on the reduced third cumulant is invoked to guarantee existence of the Fourier-Stieltjes transform, but the manuscript should explicitly verify that this transform isolates a nonzero imaginary part for a canonical oriented branching kernel (e.g., an exponential offspring density) without additional parameter tuning.

minor comments (2)

Notation for the factorial bispectrum versus the ordinary bispectrum should be clarified in the introduction to avoid reader confusion with classical higher-order spectra.
The finite-window third-order contrast in the final section would benefit from a short numerical illustration with simulated data to show practical detectability.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful reading, positive assessment, and constructive suggestion. We address the major comment below and will revise the manuscript to incorporate the requested explicit verification.

read point-by-point responses

Referee: The L^1 integrability condition on the reduced third cumulant is invoked to guarantee existence of the Fourier-Stieltjes transform, but the manuscript should explicitly verify that this transform isolates a nonzero imaginary part for a canonical oriented branching kernel (e.g., an exponential offspring density) without additional parameter tuning.

Authors: We agree that an explicit verification for a canonical kernel strengthens the presentation. In the revised manuscript we will add a dedicated example that substitutes the exponential offspring density into the general Fourier-Stieltjes transform of the reduced third cumulant. The resulting imaginary component is strictly nonzero for every positive rate parameter, with no additional tuning required. The calculation follows directly from the integral formula already derived in Section 3 and confirms that the imaginary factorial bispectrum certifies orientation. We thank the referee for this helpful suggestion. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation self-contained from third cumulant

full rationale

The paper constructs reversible nulls that match intensity and the full Bartlett spectrum, then derives the Fourier-Stieltjes transform of the reduced third cumulant for stationary Poisson branching clusters. In the L1 third-cumulant regime the nonzero imaginary factorial bispectrum is shown to certify orientation by direct computation from the process cumulants, without any fitted parameter being relabeled as a prediction, without self-citation load-bearing the central step, and without the result reducing to its own inputs by definition. The argument remains independent of the reversible nulls once the third-order transform is obtained.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The work rests on standard properties of stationary Poisson processes and cumulant measures; no new free parameters or invented entities are introduced in the abstract.

axioms (2)

standard math Stationary Poisson branching cluster processes possess well-defined reduced cumulant measures of all orders.
Invoked to define the factorial bispectrum and its Fourier-Stieltjes transform.
domain assumption The L1 integrability condition on the third cumulant guarantees the existence of the bispectrum.
Required for the nonzero imaginary part to certify orientation.

pith-pipeline@v0.9.0 · 5394 in / 1240 out tokens · 23229 ms · 2026-05-14T20:34:31.716093+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

?

unclear
Relation between the paper passage and the cited Recognition theorem.

For stationary Poisson branching clusters, we derive the Fourier–Stieltjes transform of the reduced third cumulant and show that, in the L1 third-cumulant regime, a nonzero imaginary factorial bispectrum certifies orientation.
IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Theorem 2.2. (Frequency-domain formula for the reduced complete third cumulant.) ... Bcomp(ω1,ω2)=λ R(ω1)R(ω2)R(ω3){R(−ω1)+R(−ω2)+R(−ω3)−2}

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

25 extracted references · 25 canonical work pages · 1 internal anchor

[1]

and Muzy, J.-F

Achab, M., Bacry, E., Gaïffas, S., Mastromatteo, I. and Muzy, J.-F. (2018). Uncovering causality from multivariate Hawkes integrated cumulants.Journal of Machine Learning Research18(192), 1–28

work page 2018
[2]

and Muzy, J.-F

Bacry, E. and Muzy, J.-F. (2016). First- and second-order statistics characterization of Hawkes processes and non-parametric estimation.IEEE Transactions on Information Theory62, 2184–2202

work page 2016
[3]

M., Hazelton, M

Baddeley, A., Davies, T. M., Hazelton, M. L., Rakshit, S. and Turner, R. (2022). Fundamental problems in fitting spatial cluster process models.Spatial Statistics52, 100709

work page 2022
[4]

H., Goldie, C

Bingham, N. H., Goldie, C. M. and Teugels, J. L. (1987).Regular Variation. Cambridge University Press, Cambridge

work page 1987
[5]

Brillinger, D. R. (1975). Statistical inference for stationary point processes. In M. L. Puri (ed.),Stochastic Processes and Related Topics, Vol. 1. Academic Press, New York, pp. 55–99

work page 1975
[6]

Spectral estimation of Hawkes processes from count data.The Annals of Statistics50, 1722–1746

Cheysson, Felix and Lang, Gabriel (2022). Spectral estimation of Hawkes processes from count data.The Annals of Statistics50, 1722–1746

work page 2022
[7]

Clark, D. E. (2026). Bartlett identities for spatial point processes.Statistics & Probability Letters236, 110779

work page 2026
[8]

and Toke, I

Cordi, M., Challet, D. and Toke, I. M. (2018). Testing the causality of Hawkes processes with time reversal.Journal of Statistical Mechanics: Theory and Experiment2018, 033408

work page 2018
[9]

Daley, D. J. and Vere-Jones, D. (2003).An Introduction to the Theory of Point Processes, Volume I: Elementary Theory and Methods. Springer, New York, NY

work page 2003
[10]

Dwass, M. (1969). The total progeny in a branching process and a related random walk.Journal of Applied Probability6, 682–686

work page 1969
[11]

and Dueck, J

Eichler, M., Dahlhaus, R. and Dueck, J. (2017). Graphical modeling for multivariate Hawkes processes with nonparametric link functions.Journal of Time Series Analysis38, 225–242

work page 2017
[12]

Hawkes, A. G. (1971). Spectra of some self-exciting and mutually exciting point processes.Biometrika58, 83–90

work page 1971
[13]

Hawkes, A. G. and Oakes, D. (1974). A cluster process representation of a self-exciting process.Journal of Applied Probability11, 493–503

work page 1974
[14]

Hinich, M. J. and Rothman, P. (1998). Frequency-domain test of time reversibility.Macroeconomic Dynamics2, 72–88

work page 1998
[15]

and Terdik, G

Iglói, E. and Terdik, G. (2014). When the bispectrum is real-valued.Statistics & Probability Letters95, 1–5

work page 2014
[16]

and Rotter, S

Jovanovi´c, S., Hertz, J. and Rotter, S. (2015). Cumulants of Hawkes point processes.Physical Review E91, 042802

work page 2015
[17]

Kirchner, M. (2026). Palm versions of Hawkes processes.Statistics & Probability Letters227, 110531

work page 2026
[18]

and Loisel, S

Lefèvre, C. and Loisel, S. (2013). On multiply monotone distributions, continuous or discrete, with applications. Journal of Applied Probability50, 827–847

work page 2013
[19]

Pitman, E. J. G. (1968). On the behaviour of the characteristic function of a probability distribution in the neighbourhood of the origin.Journal of the Australian Mathematical Society8, 423–443

work page 1968
[20]

Pitman, E. J. G. and Pitman, J. (2016). A direct approach to the stable distributions.Advances in Applied Probability 48(A), 261–282

work page 2016
[21]

Privault, N. (2021). Recursive computation of the Hawkes cumulants.Statistics & Probability Letters177, 109161

work page 2021
[22]

and Yamamoto, M

Takahashi, T., Nakao, M., Grüneis, F., Mizutani, Y. and Yamamoto, M. (1998). Higher-order spectra of cluster point processes generating 1/f fluctuations.Interdisciplinary Information Sciences4, 51–64

work page 1998
[23]

Spectral analysis of multivariate stationary Hawkes processes

Tang, Y., Kresin, C., Baeumer, B. and Wang, T. (2026). Spectral analysis of multivariate stationary Hawkes processes. arXiv preprint arXiv:2604.10376

work page internal anchor Pith review Pith/arXiv arXiv 2026
[24]

Williamson, R. E. (1956). Multiply monotone functions and their Laplace transforms.Duke Mathematical Journal 23, 189–207

work page 1956
[25]

and Vere-Jones, D

Zhuang, J., Ogata, Y. and Vere-Jones, D. (2002). Stochastic declustering of space-time earthquake occurrences. Journal of the American Statistical Association97, 369–380. Appendix A. Proofs and technical remarks A.1. Proof of Lemma 2.1. On any domain where the displayed logarithms are finite, the two generating-functional identities follow from the expone...

work page 2002