A singular SDE driven by additive fractional Brownian motion with Hurst parameter H<1/2

Alexander Tortoriello; Xiaoming Song

arxiv: 2604.10266 · v1 · submitted 2026-04-11 · 🧮 math.PR

A singular SDE driven by additive fractional Brownian motion with Hurst parameter H<1/2

Xiaoming Song , Alexander Tortoriello This is my paper

Pith reviewed 2026-05-10 15:35 UTC · model grok-4.3

classification 🧮 math.PR

keywords singular stochastic differential equationsfractional Brownian motionHurst parameterapproximating processesweak solutionstrajectory properties

0 comments

The pith

Solutions to singular SDEs driven by additive fractional Brownian motion with H less than 1/2 are constructed as limits of approximating processes.

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

The paper constructs solutions to a class of singular stochastic differential equations driven by additive fractional Brownian motion when the Hurst parameter satisfies H less than 1/2. It does this by introducing a family of approximating processes with smoother drivers and showing that they converge in a suitable topology to a limit that satisfies the original equation in a weak sense. A reader would care because these equations model systems with long-range dependence or rough noise where classical stochastic calculus does not apply directly. The construction also yields information on the regularity and other trajectory properties of the resulting paths. This approach therefore extends existence results from smoother regimes to the more singular case H less than 1/2.

Core claim

The solution to the singular SDE is constructed as the limit of a family of approximating processes, and its trajectory properties are investigated.

What carries the argument

A family of approximating processes whose limit satisfies the singular SDE in an appropriate weak topology.

If this is right

Existence of weak solutions is obtained for the singular equation when the driving noise is additive fractional Brownian motion with H less than 1/2.
Regularity properties of the solution paths follow from the convergence of the approximations.
The weak formulation allows the singular term to be interpreted through the limiting procedure rather than through direct stochastic integration.
Trajectory properties such as continuity or Holder continuity can be read off from the approximating family.

Where Pith is reading between the lines

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

The same limiting procedure may extend to other singular drivers whose approximations admit uniform moment bounds.
Pathwise properties established for the limit could be used to study long-time behavior or ergodicity questions for these equations.
Comparison with the H greater than 1/2 regime suggests that the singularity threshold at 1/2 is handled precisely by the choice of approximating sequence.

Load-bearing premise

The family of approximating processes converges in a suitable topology to a limit that satisfies the singular SDE in an appropriate weak sense.

What would settle it

An explicit computation or simulation in which the approximating processes fail to converge or the candidate limit fails to satisfy the integral form of the singular equation would falsify the claim.

read the original abstract

In this article we study a class of singular stochastic differential equations driven by fractional Brownian motion with Hurst parameter H<1/2. The solution is constructed as the limit of a family of approximating processes, and its trajectory properties are investigated.

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 builds solutions to singular SDEs with additive fBM for H<1/2 by approximation and limit, but the step that passes the singular drift to the limit is not obviously closed.

read the letter

The core contribution is an approximation scheme that produces a candidate solution to the singular equation and then checks its path properties. This targets the H<1/2 regime, which has seen less work than the smoother case, so the extension itself is the main new piece. The setup looks standard for this subfield: regularize either the coefficient or the noise, obtain uniform estimates, and pass to the limit in a suitable topology. Trajectory regularity follows from the driving noise once the limit exists. That part is routine but useful to have written down for this parameter range. The soft spot is exactly where the stress-test note points. Even with convergence of the processes, the integral of b against the approximants need not converge to the integral against the limit when b is discontinuous and the paths are only Hölder-H with H<1/2. Without an explicit identification argument, a monotonicity condition, or a weak formulation that lets one test against smooth functions, the limit may solve a different equation. The abstract gives no hint that such a step is supplied, so the central claim rests on an unverified passage. This is a genuine gap rather than a minor technicality. The work is aimed at specialists in stochastic analysis who already know the H>1/2 literature and want to see the rougher case handled. A reader outside that niche will find it too narrow. I would send it to referees because the question is well-posed and the approximation approach is reasonable; the referees can check whether the limit identification actually holds. If it does, the paper is worth having; if not, it needs a different formulation or extra assumptions on b.

Referee Report

2 major / 2 minor

Summary. The manuscript constructs solutions to a class of singular SDEs driven by additive fractional Brownian motion with Hurst index H < 1/2. The solution is obtained as the limit of a family of approximating processes, after which the trajectory regularity and other path properties of the limiting process are studied.

Significance. If the limit identification is rigorously established, the result would extend the theory of SDEs with rough additive noise to the singular-drift regime for H < 1/2, where standard Young or rough-path integration is unavailable and the low Hölder regularity of the driver makes passage to the limit delicate.

major comments (2)

[§3] §3 (Construction of the limit): the argument that X_n → X implies ∫ b(X_n(s)) ds → ∫ b(X(s)) ds for discontinuous or unbounded b is not made explicit. With only Hölder-H paths for H < 1/2, uniform convergence alone does not guarantee the nonlinear term passes to the limit; a weak formulation, monotonicity assumption on b, or explicit identification step is required.
[Theorem 4.1] Theorem 4.1 (or equivalent statement of the main existence result): the topology of convergence for the approximating family is not stated with sufficient precision to control the singular drift integral, leaving open whether the limit satisfies the original SDE in the intended sense.

minor comments (2)

[Abstract] The abstract omits any mention of the assumptions placed on the drift b (e.g., continuity, growth, or monotonicity conditions), which are essential for the construction.
[§2] Notation for the approximating processes and the limiting integral equation should be introduced earlier and used consistently throughout §2–§4.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and the insightful comments on the limit construction and convergence topology. We address each major point below and will revise the manuscript to make the arguments more explicit.

read point-by-point responses

Referee: [§3] §3 (Construction of the limit): the argument that X_n → X implies ∫ b(X_n(s)) ds → ∫ b(X(s)) ds for discontinuous or unbounded b is not made explicit. With only Hölder-H paths for H < 1/2, uniform convergence alone does not guarantee the nonlinear term passes to the limit; a weak formulation, monotonicity assumption on b, or explicit identification step is required.

Authors: We agree that the passage to the limit in the drift integral requires a more explicit justification, particularly given the possible lack of Lipschitz regularity or boundedness in b. In the construction, X_n converges almost surely to X uniformly on compact time intervals (as a consequence of the Hölder-α convergence for α < H). Under the standing assumptions that b is continuous with at most polynomial growth, pointwise convergence of b(X_n(s)) to b(X(s)) combined with uniform integrability (from moment bounds on the paths) yields the desired convergence of the integrals via the dominated convergence theorem. We acknowledge that this step was not written out in full detail. In the revised manuscript we will insert a dedicated lemma in §3 that spells out the precise conditions on b and verifies the limit passage explicitly. If the manuscript is intended to cover discontinuous b, we will restrict the statement or add a weak-convergence argument as needed. revision: yes
Referee: [Theorem 4.1] Theorem 4.1 (or equivalent statement of the main existence result): the topology of convergence for the approximating family is not stated with sufficient precision to control the singular drift integral, leaving open whether the limit satisfies the original SDE in the intended sense.

Authors: We will revise the statement of the main existence result (Theorem 4.1 and the surrounding discussion in §3) to specify the topology explicitly: the approximating processes converge almost surely to the limit in the Hölder space C^α([0,T]) for every α < H. This topology is stronger than uniform convergence and directly controls the integrals against continuous functions of at most polynomial growth. We will also add a short paragraph explaining why this mode of convergence ensures that the limiting process satisfies the original integral equation with the singular drift term. revision: yes

Circularity Check

0 steps flagged

No circularity: standard limit-of-approximations construction for singular SDE

full rationale

The paper constructs the solution explicitly as the limit of a family of approximating processes in a suitable topology and then studies trajectory properties of that limit. This is a conventional existence strategy for singular SDEs and does not reduce any claimed result to a self-definition, a fitted parameter renamed as a prediction, or a load-bearing self-citation. No equations are supplied that equate the target object to its own inputs by construction, and the abstract gives no indication that the identification of the limit equation is smuggled in via prior work by the same authors. The derivation is therefore self-contained against external convergence arguments.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no information on free parameters, background axioms, or newly postulated entities.

pith-pipeline@v0.9.0 · 5321 in / 924 out tokens · 59161 ms · 2026-05-10T15:35:47.596936+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

12 extracted references · 12 canonical work pages

[1]

Boufoussi and Y

B. Boufoussi and Y. Ouknine. On a SDE driven by a fractional Brownian motion and with monotone drift.Electron. Comm. Probab., 8:122–134, 2003

work page 2003
[2]

E. H. Essaky and D. Nualart. On the 1 H -variation of the divergence integral with respect to fractional Brownian motion with Hurst parameterH < 1 2. Stochastic Process. Appl., 125(11):4117–4141, 2015

work page 2015
[3]

A. M. Garsia, E. Rodemich, and H. Rumsey, Jr. A real variable lemma and the continuity of paths of some Gaussian processes.Indiana Univ. Math. J., 20:565–578, 1970/71

work page 1970
[4]

Jo˜ ao M. E. Guerra and David Nualart. The 1/H-variation of the divergence integral with respect to the fractional Brownian motion forH >1/2 and fractional Bessel processes.Stochastic Process. Appl., 115(1):91–115, 2005

work page 2005
[5]

Hu and D

Y. Hu and D. Nualart. Some processes associated with fractional Bessel processes.J. Theoret. Probab., 18(2):377–397, 2005

work page 2005
[6]

Y. Hu, D. Nualart, and X. Song. A singular stochastic differential equation driven by fractional Brownian motion.Statist. Probab. Lett., 78(14):2075– 2085, 2008. 19

work page 2075
[7]

Karatzas and S

I. Karatzas and S. E. Shreve.Brownian motion and stochastic calculus, volume 113 ofGraduate Texts in Mathematics. Springer-Verlag, New York, second edition, 1991

work page 1991
[8]

Mishura and D

Y. Mishura and D. Nualart. Weak solutions for stochastic differential equa- tions with additive fractional noise.Statist. Probab. Lett., 70(4):253–261, 2004

work page 2004
[9]

Mishura and A

Y. Mishura and A. Yurchenko-Tytarenko. Fractional Cox-Ingersoll-Ross process with small Hurst indices.Mod. Stoch. Theory Appl., 6(1):13–19, 2019

work page 2019
[10]

Nualart.The Malliavin calculus and related topics

D. Nualart.The Malliavin calculus and related topics. Probability and its Applications (New York). Springer-Verlag, Berlin, second edition, 2006

work page 2006
[11]

Nualart and Y

D. Nualart and Y. Ouknine. Regularization of differential equations by fractional noise.Stochastic Process. Appl., 102(1):103–116, 2002

work page 2002
[12]

Nualart and A

D. Nualart and A. R˘ a¸ scanu. Differential equations driven by fractional Brownian motion.Collect. Math., 53(1):55–81, 2002. 20

work page 2002

[1] [1]

Boufoussi and Y

B. Boufoussi and Y. Ouknine. On a SDE driven by a fractional Brownian motion and with monotone drift.Electron. Comm. Probab., 8:122–134, 2003

work page 2003

[2] [2]

E. H. Essaky and D. Nualart. On the 1 H -variation of the divergence integral with respect to fractional Brownian motion with Hurst parameterH < 1 2. Stochastic Process. Appl., 125(11):4117–4141, 2015

work page 2015

[3] [3]

A. M. Garsia, E. Rodemich, and H. Rumsey, Jr. A real variable lemma and the continuity of paths of some Gaussian processes.Indiana Univ. Math. J., 20:565–578, 1970/71

work page 1970

[4] [4]

Jo˜ ao M. E. Guerra and David Nualart. The 1/H-variation of the divergence integral with respect to the fractional Brownian motion forH >1/2 and fractional Bessel processes.Stochastic Process. Appl., 115(1):91–115, 2005

work page 2005

[5] [5]

Hu and D

Y. Hu and D. Nualart. Some processes associated with fractional Bessel processes.J. Theoret. Probab., 18(2):377–397, 2005

work page 2005

[6] [6]

Y. Hu, D. Nualart, and X. Song. A singular stochastic differential equation driven by fractional Brownian motion.Statist. Probab. Lett., 78(14):2075– 2085, 2008. 19

work page 2075

[7] [7]

Karatzas and S

I. Karatzas and S. E. Shreve.Brownian motion and stochastic calculus, volume 113 ofGraduate Texts in Mathematics. Springer-Verlag, New York, second edition, 1991

work page 1991

[8] [8]

Mishura and D

Y. Mishura and D. Nualart. Weak solutions for stochastic differential equa- tions with additive fractional noise.Statist. Probab. Lett., 70(4):253–261, 2004

work page 2004

[9] [9]

Mishura and A

Y. Mishura and A. Yurchenko-Tytarenko. Fractional Cox-Ingersoll-Ross process with small Hurst indices.Mod. Stoch. Theory Appl., 6(1):13–19, 2019

work page 2019

[10] [10]

Nualart.The Malliavin calculus and related topics

D. Nualart.The Malliavin calculus and related topics. Probability and its Applications (New York). Springer-Verlag, Berlin, second edition, 2006

work page 2006

[11] [11]

Nualart and Y

D. Nualart and Y. Ouknine. Regularization of differential equations by fractional noise.Stochastic Process. Appl., 102(1):103–116, 2002

work page 2002

[12] [12]

Nualart and A

D. Nualart and A. R˘ a¸ scanu. Differential equations driven by fractional Brownian motion.Collect. Math., 53(1):55–81, 2002. 20

work page 2002