pith. sign in

arxiv: 1906.09813 · v1 · pith:ZCGORYOOnew · submitted 2019-06-24 · 🧮 math.PR

Simulation of Conditioned Diffusions on the Flat Torus

Mathias H{\o}jgaard Jensen , Anton Mallasto , Stefan Sommer This is my paper

Pith reviewed 2026-05-25 17:25 UTC · model grok-4.3

classification 🧮 math.PR
keywords conditioned diffusionsflat torussimulationconvergenceBrownian motionprojectionstochastic processestoroidal geometry
0
0 comments X

The pith

A diffusion in the plane, conditioned to a terminal point, projects onto the flat torus and converges to that point.

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

The paper develops a simulation method for diffusions on the flat torus that are conditioned to reach a fixed terminal point after a set time. The approach runs an ordinary diffusion in the covering space R² and projects the paths down to the torus. It proves that the projected paths converge to the target point on the torus. The same construction yields a local equivalence to Brownian motion after an appropriate change of measure. This supplies a practical way to generate conditioned paths on toroidal geometries without extra restrictions on the driving noise or the time horizon.

Core claim

Consider a diffusion process in R² conditioned on hitting a prescribed point at a fixed terminal time. Its projection onto the flat torus converges to the corresponding terminal point on the torus. Under a suitable change of measure the Euclidean diffusion is locally a Brownian motion.

What carries the argument

The projection map from R² onto the flat torus that carries the terminal conditioning from the plane to the torus.

If this is right

  • Paths generated on the torus reach the prescribed terminal point with probability approaching one.
  • The simulation requires no additional constraints on the driving noise or the length of the time interval.
  • The local Brownian-motion property under the changed measure simplifies sampling and analysis of the paths.
  • The method directly supplies conditioned trajectories for stochastic models on toroidal domains.

Where Pith is reading between the lines

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

  • Similar lifting-and-projection arguments might apply to conditioned processes on other compact flat manifolds.
  • The technique could be tested numerically by comparing hitting probabilities on the torus against direct toroidal simulations for moderate time horizons.
  • In applications the local Brownian property may allow reuse of existing Euclidean samplers with only a measure adjustment.
  • Extensions to non-flat tori would require checking whether the projection still preserves conditioning after curvature is introduced.

Load-bearing premise

The projection of the conditioned diffusion from R² onto the torus preserves the terminal conditioning and produces the stated convergence without further restrictions on the driving process or time horizon.

What would settle it

A concrete diffusion in R² whose projection onto the torus fails to approach the target point in probability as the horizon is reached, or an explicit calculation showing that the change-of-measure property does not hold locally.

Figures

Figures reproduced from arXiv: 1906.09813 by Anton Mallasto, Mathias H{\o}jgaard Jensen, Stefan Sommer.

Figure 1
Figure 1. Figure 1: The figure illustrates the possibility of the diffusion path going an arbi [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Two different paths visualized both on the torus and in Eucliden space. [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Figure 3a depicts the evolution of the drift term. It shows how the pull [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Figure 4 shows 9 paths from the proposed model (4) on the left and the [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
read the original abstract

Diffusion processes are fundamental in modelling stochastic dynamics in natural sciences. Recently, simulating such processes on complicated geometries has found applications for example in biology, where toroidal data arises naturally when studying the backbone of protein sequences, creating a demand for efficient sampling methods. In this paper, we propose a method for simulating diffusions on the flat torus, conditioned on hitting a terminal point after a fixed time, by considering a diffusion process in R 2 which we project onto the torus. We contribute a convergence result for this diffusion process, translating into convergence of the projected process to the terminal point on the torus. We also show that under a suitable change of measure, the Euclidean diffusion is locally a Brownian motion.

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

1 major / 0 minor

Summary. The manuscript proposes simulating conditioned diffusions on the flat torus T^2 = R^2/Z^2 by defining a conditioned diffusion in R^2 and projecting it onto the torus. It claims a convergence result showing that the projected process converges to the prescribed terminal point on T^2, and that under a suitable change of measure the Euclidean diffusion is locally a Brownian motion.

Significance. If the convergence and change-of-measure claims hold with the necessary conditions made explicit, the approach could provide an efficient simulation method for conditioned paths on toroidal geometries, relevant to applications such as modeling protein backbones in biology. The local Brownian-motion property under reweighting would be a useful simplification for sampling.

major comments (1)
  1. [Abstract] Abstract: the convergence claim for the projected process rests on the assertion that conditioning in R^2 to a single point x_T and then projecting yields convergence to the terminal point on T^2. However, the terminal condition on T^2 corresponds to the entire lattice x_T + Z^2; the abstract gives no indication that the construction accounts for the mixture over lifts or imposes restrictions (e.g., small T or initial distribution supported in a fundamental domain) that would make the projected law identical to the conditioned law on T^2. This is load-bearing for the central claim.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their careful reading of the manuscript and for highlighting an important point about the abstract and the central convergence claim. We respond to the major comment below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the convergence claim for the projected process rests on the assertion that conditioning in R^2 to a single point x_T and then projecting yields convergence to the terminal point on T^2. However, the terminal condition on T^2 corresponds to the entire lattice x_T + Z^2; the abstract gives no indication that the construction accounts for the mixture over lifts or imposes restrictions (e.g., small T or initial distribution supported in a fundamental domain) that would make the projected law identical to the conditioned law on T^2. This is load-bearing for the central claim.

    Authors: We agree that the abstract does not explicitly address the mixture over lattice lifts or the conditions needed for the projected process to have exactly the same law as a conditioned diffusion on the torus. The convergence result established in the paper is that the Euclidean process converges to the chosen lift x_T, which implies that its projection converges to the corresponding point on T^2. However, to ensure the finite-dimensional distributions of the projected process match those of the torus-conditioned diffusion, the construction implicitly relies on the probability of hitting other lifts being negligible (e.g., for small T or when the initial distribution is supported in a fundamental domain). We will revise the abstract to state these assumptions explicitly and add a clarifying remark in the introduction and in the statement of the main convergence theorem. revision: yes

Circularity Check

0 steps flagged

No circularity: convergence claim derived from standard projection and diffusion theory

full rationale

The paper's central contribution is a convergence result obtained by lifting the conditioned diffusion to R^2, projecting to the torus, and invoking standard diffusion convergence under the projection map. No equation reduces a claimed prediction to a fitted parameter by construction, no uniqueness theorem is imported via self-citation, and no ansatz is smuggled in. The derivation chain remains self-contained against external benchmarks of conditioned Brownian motion and torus quotients; the reader's score of 2 reflects only routine self-citation of background results that are independently verifiable.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard properties of diffusions and Brownian motion in Euclidean space together with the unstated technical conditions under which the projection and change of measure preserve the conditioning.

axioms (1)
  • standard math Standard properties of diffusion processes and Brownian motion in Euclidean space
    The projection method and local Brownian motion claim presuppose the usual Ito calculus and measure-theoretic foundations for diffusions.

pith-pipeline@v0.9.0 · 5644 in / 1116 out tokens · 37244 ms · 2026-05-25T17:25:12.825466+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.

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

13 extracted references · 13 canonical work pages

  1. [1]

    Stochastic Processes and their Applications 116(11), 1660--1675 (2006)

    Delyon, B., Hu, Y.: Simulation of conditioned diffusion and application to parameter estimation. Stochastic Processes and their Applications 116(11), 1660--1675 (2006)

    work page 2006

  2. [2]

    Springer (1989)

    Emery, M.: Stochastic calculus in manifolds. Springer (1989)

    work page 1989

  3. [3]

    Statistics and Computing pp

    Garc \' a-Portugu \'e s, E., S rensen, M., Mardia, K.V., Hamelryck, T.: Langevin diffusions on the torus: estimation and applications. Statistics and Computing pp. 1--22 (2017)

    work page 2017

  4. [4]

    Molecular Biology and Evolution 34(8), 2085--2100 (2017)

    Golden, M., García-Portugués, E., Sørensen, M., Mardia, K.V., Hamelryck, T., Hein, J.: A Generative Angular Model of Protein Structure Evolution . Molecular Biology and Evolution 34(8), 2085--2100 (2017)

    work page 2085

  5. [5]

    Hsu, E.P.: Stochastic analysis on manifolds, vol. 38. American Mathematical Soc. (2002)

    work page 2002

  6. [6]

    Springer, New York, 2nd edition edn

    Karatzas, I., Shreve, S.: Brownian Motion and Stochastic Calculus . Springer, New York, 2nd edition edn. (Aug 1991)

    work page 1991

  7. [7]

    Theory of Probability & Its Applications 17(4), 717--720 (Sep 1973)

    Novikov, A.A.: On an Identity for Stochastic Integrals . Theory of Probability & Its Applications 17(4), 717--720 (Sep 1973)

    work page 1973

  8. [8]

    Princeton University Press (Mar 2014)

    Sogge, C.D.: Hangzhou Lectures on Eigenfunctions of the Laplacian ( AM -188). Princeton University Press (Mar 2014)

    work page 2014

  9. [9]

    In: Graphs in Biomedical Image Analysis, Computational Anatomy and Imaging Genetics, pp

    Sommer, S., Arnaudon, A., Kuhnel, L., Joshi, S.: Bridge simulation and metric estimation on landmark manifolds. In: Graphs in Biomedical Image Analysis, Computational Anatomy and Imaging Genetics, pp. 79--91. Springer (2017)

    work page 2017

  10. [10]

    Theory of Probability & Its Applications 24(2), 354--366 (1980)

    Veretennikov, A.Y.: On the strong solutions of stochastic differential equations. Theory of Probability & Its Applications 24(2), 354--366 (1980)

    work page 1980

  11. [11]

    Sbornik: Mathematics 39, 387--403 (1981)

    Veretennikov, A.J.: On strong solutions and explicit formulas for solutions of stochastic integral equations. Sbornik: Mathematics 39, 387--403 (1981)

    work page 1981

  12. [12]

    , " * write output.state after.block = add.period write

    ENTRY address author booktitle chapter doi edition editor eid howpublished institution journal key month note number organization pages publisher school series title type url volume year label INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.all := #1 'mid.sentence := #2 'after.sentence := #3 '...

    work page

  13. [13]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize ":" * " " *...

    work page