Well-posedness for Wright--Fisher SPDEs with measurable drift
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: I'm Ada.
Emmy: And I'm Emmy, here with Kolya.
Ada: We're looking at the paper: "Well-posedness for Wright--Fisher SPDEs with measurable drift".
Emmy: Here's the gist of the paper. We prove weak existence and uniqueness in law of
0, 1: -valued solutions for a class of scalar stochastic heat equations with measurable drifts b:
0, 1: → R, driven by Wright–Fisher noise The only assumptions we require on the drift are Borel-measurability, and the one-way inequalities −Cu ≤ b(u) ≤ C(1−u) for some C < ∞, compatible with the necessary condition b(1) ≤ 0 ≤ b(0).
Ada: Our proof relies on a stochastic duality between the solution of the stochastic heat equation, and a voting scheme running along the graph of a branching-coalescing particle system, generalising the class of drifts for which a dual process is available.
Emmy: Problem Setup and Key Assumptions The paper considers the one-dimensional SPDE ∂tu = 1/2 ∆u + b(u) + p ν u(1 − u)W, defined on a connected spatial domain E ⊆ R with initial condition u0: E →
0, 1: (assumed only to be a measurable function), where b:
0, 1: → R is a bounded, measurable function satisfying −Cu ≤ b(u) ≤ C(1−u) for some C < ∞.
Ada: The well-posedness of
0, 1: -valued solutions requires that b(0) ≥ 0 and b(1) ≤ 0, otherwise the solutions could be “pushed outside of [0, 1)”.
Paper summary: Ada: So, to recap, the core thesis of this paper is that you can have weak existence and uniqueness in law for solutions of the stochastic heat equation when the drift function is only Borel measurable >
Emmy: They establish this by setting up a duality between the SPDE solution and a branching-coalescing particle system, which they use to determine the moments of that solution, which ultimately defines its law >
Kolya: But what's really interesting for me is how they extend the work on regularization by noise. They take that result from Barnes, Mytnik and Sun and apply it to this much broader class of drifts >
Ada: Right. The paper says their technique allows them to handle non-differentiable drift functions using a voting scheme based on a Bernoulli factory for the branching mechanism >
Emmy: That sounds like they're creating a way to handle the discontinuities in b by using this particle system structure, which is where my geometer intuition kicks in, because that duality is really mapping continuous structures onto discrete ones >
Kolya: From a probabilistic standpoint, I’m thinking about what the bounds on these moments actually tell us about the tail behavior of the solution distribution when you drive it with Wright--Fisher noise >
Ada: And they show that these moment duals converge under pointwise limits of the drifts, meaning if you have a sequence of drifts b n approaching some drift b, then the corresponding solutions u n approach u in law >
Emmy: That convergence in distribution is what allows them to establish uniqueness in law by showing that the moment duals converge as well >
Kolya: So they're saying that even if you have these rough, measurable drifts, as long as they satisfy those one-way bounds, the resulting probability distributions for the solutions are uniquely determined >
Ada: That’s the main point: it takes a lot of technical machinery and a clever duality argument to get weak existence and uniqueness in law for this general class of drift functions >
Conclusion: Ada: Looking at the title, "Well-posedness for Wright--Fisher SPDEs with measurable drift," it tells you exactly what this work is about: extending our understanding of these specific type of equations to handle drifts that aren't perfectly smooth >
Emmy: It seems the authors are saying that we don't need those very restrictive conditions on the drift function for solutions to even exist in a weak sense >
Kolya: So, if you take this result and apply it to real-world problems where noise or external influences might lead to non-smooth dynamics, this paper gives us more tools to make predictions about what the system will look like >
Ada: It really does. The implication is that we can study these processes under much more realistic conditions than just assuming the drift is continuous >
Emmy: I think it means we gain a robust method for analyzing systems where the governing rules, represented by b, have some jumps or abrupt changes along the interval zero to one >
Kolya: The caveat they mention is that they still need those strict bounds on the drift function, and if you break those bounds, well-posedness might fall apart >
Ada: Right. So it’s a very powerful result because it broadens the scope of what we can study for these equations while keeping the mathematical structure sound >
Emmy: It’s about showing that the underlying algebraic geometry of this duality holds even when things get messy with measurable functions >
Jere Koskela, Oliver Tough
Graz University of Technology · Durham University
math.PR, math.AP
Submitted: 2026-09-20
Updated: 2026-11-13
Comments: 27 pages, v2 plugs a gap in the proof in v1
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 91/100
The gist: We prove weak existence and uniqueness in law of [0, 1]-valued solutions for a class of scalar stochastic heat equations with measurable drifts b: [0, 1] → R…
Key concepts
- Stochastic Heat Equation (SPDE)
- This is a partial differential equation where one of the terms involves randomness (noise). It models how some quantity, like concentration or population density, evolves over space and time. The specific equation here includes a drift term $b(u)$ that depends on the current state $u$.
- Wright–Fisher Noise
- This is a specific type of noise used in the SPDE. It models stochastic fluctuations often seen in population genetics, where individuals reproduce or die based on random events. It drives the randomness in the system's evolution.
- Stochastic Duality
- This is a powerful mathematical technique used to relate two different problems—in this case, the SPDE and a particle system. By finding this link, researchers can use known results from one problem to solve or understand the other, which helps establish existence and uniqueness.
Terminology
Summary
We prove weak existence and uniqueness in law of [0, 1]-valued solutions for a class of scalar stochastic heat equations with measurable drifts b: [0, 1] → R, driven by Wright–Fisher noise The only assumptions we require on the drift are Borel-measurability, and the one-way inequalities −Cu ≤ b(u) ≤ C(1−u) for some C < ∞, compatible with the necessary condition b(1) ≤ 0 ≤ b(0). p. 0 uk
Department of Mathematical Sciences,
Durham Univeristy
September 22, 2026
Abstract
We prove weak existence and uniqueness in law of [0, 1]-valued solutions for a class of
scalar stochastic heat equations with measurable drifts b : [0, 1] → R, driven by Wright–
Fisher noise. The only assumptions we require on the drift are Borel-measurability, and the
one-way inequalities −Cu ≤ b(u) ≤ C(1−u) for some C < ∞, compatible with the necessary
condition b(1) ≤ 0 ≤ b(0). These are vastly more general conditions than were previously
available. Our proof relies on a stochastic duality between the solution of the stochastic heat
equation, and a voting scheme running along the graph of a branching-coalescing particle
system, generalising the class of drifts for which a dual process is available.Where the paper says it
Where the paper says it
Problem Setup and Key Assumptions
The paper considers the one-dimensional SPDE ∂tu = 1/2 ∆u + b(u) + p ν u(1 − u)W, defined on a connected spatial domain E ⊆ R with initial condition u0: E → [0, 1] (assumed only to be a measurable function), where b: [0, 1] → R is a bounded, measurable function satisfying −Cu ≤ b(u) ≤ C(1−u) for some C < ∞. The well-posedness of [0, 1]-valued solutions requires that b(0) ≥ 0 and b(1) ≤ 0, otherwise the solutions could be “pushed outside of [0, 1)”. p. 0 Specifically,
we will assume that b : [0, 1] → R is a bounded, measurable function satisfying
−Cu ≤ b(u) ≤ C(1 − u) for some C < ∞. (2)
In order to ensure the well-posedness of [0, 1]-valued solutions of (1), it is necessary that b(0) ≥ 0
and b(1) ≤ 0, otherwise the solutions could be “pushed outside of [0, 1]”. Assume that b is Borel measurable and satisfies (2), ν > 0, and any one of
the boundary conditions (i)-(iv) hold. Then for any (possibly random) Borel-measurable initial
condition u0 : E → [0, 1], there exists a weak solution of (1), which is unique in law.Where the paper says it
Where the paper says it
Duality and Regularization by Noise
The proof strategy involves establishing a duality between solutions of (5) and a suitable branching-coalescing particle system. This duality is used to determine moments E[ut(x1). p. 2 Our technique will be to establish a duality between solutions of (5) and a suitable branchingcoalescing particle system. We will use that duality to determine moments E[ut(x1). We will use that duality to determine moments E[ut(x1). ut(xN )] of
the solution, which determine its law (Athreya and Tribe, 2000, Lemma 1). Our proof relies on a stochastic duality between the solution of the stochastic heat
equation, and a voting scheme running along the graph of a branching-coalescing particle
system, generalising the class of drifts for which a dual process is available. Our results
and their proof put the regularisation-by-noise result of Barnes, Mytnik and Sun, and its
explanation, on a much more general footing. (2025) in particular. We use a voting scheme based on a
Bernoulli factory as the branching mechanism for particles, which makes it possible to handle
non-differentiable drift functions.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Convergence of Solutions and Uniqueness
The paper proves convergence of solutions under pointwise limits of drifts, stating that if b n → b pointwise, then u n → u locally uniformly in distribution as n → ∞. Uniqueness in law is established by showing that the moment duals converge under these conditions. p. 2 Hence, they are natural objects with which to replace drift
terms given by branching process moment-generating functions when the branching property is
not available. We will further extend this to discontinuous drifts by showing that the value of
our duality representations converge under pointwise limits of the drifts, notwithstanding that
the dual branching process process may not itself converge as there is no Bernoulli factory for
discontinuous drifts (see Theorem 3. For any
pair of paricles, their intersection local time is equal to the local time process Lt of a single
rate 2 Brownian motion at 0 (the Brownian motion given by their difference). This local time
process is equal in law to the running maximum of a single Brownian motion started at 0, hence
Lt ≈
√
t for small t.Where the paper says it
Where the paper says it
Existence for Measurable Drifts
The existence of a solution for all Borel-measurable b satisfying −Cu ≤ b(u) ≤ C(1−u) follows by transfinite induction, leveraging Lemma 4.1. This lemma shows that if (b n) is a sequence of measurable functions, each possessing a solution u n, and b n → b pointwise, then b possesses a weak solution also. p. 1 Assume that b is Borel measurable and satisfies (2), ν > 0, and any one of
the boundary conditions (i)-(iv) hold. Then for any (possibly random) Borel-measurable initial
condition u0 : E → [0, 1], there exists a weak solution of (1), which is unique in law.Where the paper says it
Tightness and Weak Existence
To establish existence for general bounded and measurable drifts, the paper proves tightness in Cb,loc(R>0 × E; [0, 1]) using Lemma 3.3. This lemma shows that the set of all solutions u(t, x) to (5), for all possible initial conditions and all measurable drifts b: E × R → [0, 1] with b∞ ≤ C, is tight in Cb,loc(R>0 × E; [0, 1]). p. 4 We then extend this duality to Baire class-1 drifts, which we use to establish wellposedness in Section 3. Finally, we extend our proof of duality to Borel-measurable drifts in
Section 4
2 Duality for continuous drifts
Throughout this section, we will assume that b is continuous and satisfies (2), and that u is an
arbitrary weak solution of (1) for one of the chosen boundary conditions (i)-(iv), with initial
condition u0: a (possibly random) Borel-measurable function E → [0, 1].Where the paper says it
Uniqueness via Measure Convergence
Uniqueness in law is established by defining a bounded Borel measure ν on [0, 1] in terms of u 1 and u 2 as ν(·):= X i=1,2 X R∈N 2 −RE " Z R 0 Z [-R,R]∩E δui(t,x) dxdt. This measure allows for the construction of a sequence of continuous functions ˜b m converging ν-almost everywhere to ˜b. p. 5 1, we henceforth consider the one-dimensional SPDE
∂tu =
1
2
∆u + µ(
˜b(u) − u), +
p
νu(1 − u)W , ˙ (5)
where ˜b : [0, 1] → [0, 1] is assumed to satisfy (4), W˙
is space-time white noise, µ, ν > 0 are
constants, and u satisfies one of the boundary conditions (i)-(iv) described in the introduction. Since ˜b satisfies the conditions of Keane and O’Brien (1994) for the existence of a Bernoulli
factory, there exists a sequence of continuous and polynomially bounded functions {
˜bk}k≥1, as
well as positive integers {ηk}k≥1 such that
˜b(u) = X∞
k=1
3
4
k−1 1
4
P(
˜bk(Y (ηk, u)/ηk) ≥ 1/2), (6)
where Y (n, u) ∼ Bin(n, u). E = R/Z, (iii) E = [0, 1] with Dirichlet
boundary conditions, u(0) = u(1) = 0, or (iv) E = [0, 1] with Neumann boundary conditions,
∂xu(0) = ∂xu(1) = 0. We are interested in uniqueness in law of solutions to (1) when b is irregular.Where the paper says it
Where the paper says it
Existence via Approximation
The existence of weak solutions for Baire class-1 drifts is shown by approximating them with continuous and polynomially bounded functions ˜b m converging pointwise to b. The convergence of the duals under these approximations implies that u n,m → u n locally uniformly in distribution as m → ∞. p. 1 Let u
n and u be the weak solutions of (1)
with drift b
n and b, respectively. Then u
n → u locally uniformly in distribution as n → ∞. These are vastly more general conditions than were previously
available. Our proof relies on a stochastic duality between the solution of the stochastic heat
equation, and a voting scheme running along the graph of a branching-coalescing particle
system, generalising the class of drifts for which a dual process is available.Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improvement of weak existence for irregular drifts: The system can now establish
weak existence and uniqueness in law of [0, 1]-valued solutions for a class of scalar stochastic heat equations with measurable drifts b: [0, 1] → R,
even when b is only Borel-measurable and satisfies the one-way inequalities −Cu ≤ b(u) ≤ C(1−u). -
Improvement in handling discontinuous drifts: The method allows establishing
weak existence for all Borel-measurable b satisfying −Cu ≤ b(u) ≤ C(1−u),
extending results beyond those restricted to power series or analytic functions. -
Improvement in convergence analysis: The system can now prove that
the u n → u locally uniformly in distribution as n → ∞
even when the drifts b n converge pointwise, addressing difficulties caused by discontinuities of b. -
Improvement in uniqueness results: Uniqueness in law is established for a broader class of drifts, including Baire class-1 functions satisfying the one-way Lipschitz condition and being bounded and measurable.
-
Improvement in solution existence proof: The system can prove that
the existence of a solution for all Borel-measurable b satisfying −Cu ≤ b(u) ≤ C(1−u) follows by transfinite induction,
providing a complete framework for the existence of weak solutions for this class of drifts.
Abstract
We prove weak existence and uniqueness in law of [0,1] -valued solutions for a class of scalar stochastic heat equations with measurable drifts b: [0,1] to R, driven by Wright--Fisher noise. The only assumptions we require on the drift are Borel-measurability, and the one-way inequalities-Cu at most b(u) at most C(1-u) for some C< infinity, compatible with the necessary condition b(1) at most 0 at most b(0). These are vastly more general conditions than were previously available. Our proof relies on a stochastic duality between the solution of the stochastic heat equation, and a voting scheme running along the graph of a branching-coalescing particle system, generalising the class of drifts for which a dual process is available. Our results and their proof put the regularisation-by-noise result of Barnes, Mytnik and Sun, and its explanation, on a much more general footing.
Sources
- Fluctuations for fully pushed stochastic fronts
- Quasi-stationary behavior of the stochastic FKPP equation on the circle
- Weak Existence and Uniqueness for Super-Brownian Motion with Irregular Drift
Related papers
- Rank-One HCIZ Integrals and Multiplicative Convolutions
- Repeated averaging on random d-regular graphs
- Precise cover times for branching random walks on Hamming graphs: (iterated) logarithmic corrections
- Markov Chain CLTs: Resolving Open Problems
- Marked stable L'evy processes in Brownian local time
- A Sharp Refinement of Samuels' Inequality via Cap Moves