Pith. sign in

REVIEW 3 major objections 4 minor 43 references

Exponential Convergence of the Sinkhorn Algorithm for the Schr\"odinger Bridge with Regime Switching

T0 review · 3 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper proves exponential convergence of the Sinkhorn algorithm for Schrödinger bridges with regime switching in relative entropy.

desk verdict First exponential convergence guarantee for Sinkhorn in regime-switching Schrödinger bridges; core proof is sound but the main theorem rests on a regularity assumption verified only for constant coefficients. read the letter →

arxiv 2607.19176 v1 pith:PTPK6UIF submitted 2026-07-21 math.PR

Katharina Eichinger , Anna Kazeykina , Zhenjie Ren , Hecheng Wang This is my paper · ORCID
classification math.PR MSC 60J6060J2749Q22
keywords SinkhornalgorithmSchrödingerbridgeregimeswitchingdiffusionentropicoptimaltransportrelativeentropyexponentialconvergencesemiconcavityhybridstatespace
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper proves that the Sinkhorn algorithm—the standard iterative solver for entropic optimal transport—converges exponentially in relative entropy when applied to the Schrödinger bridge problem for a regime-switching diffusion, a hybrid process whose state combines a continuous position with a discrete regime label. The main result, Theorem 2.12, shows that under compact-support, smoothness, and Talagrand-type conditions the half-step and full-step Sinkhorn couplings approach the optimal bridge at a geometric rate; Theorem 2.21 extends the same guarantee to the partially observed setting where only the terminal continuous distribution is specified. A sympathetic reader should see this as extending the exponential-convergence theory of Sinkhorn from classical Euclidean spaces to a hybrid state space, where the cost function couples continuous motion and switching between regimes. The proof quantifies stability of the optimal bridge under perturbations of the terminal law, then combines that bound with a known contraction in total variation to force a recursion that decays exponentially.

What carries the argument

The central machinery consists of the Sinkhorn iteration written on the hybrid state space, with cost c_ij(x,y) = -log r_ij(0,x;T,y) built from the reference transition density. The proof carries three load-bearing components: (i) a semiconcavity estimate (Lemma 3.4) showing that the cost plus the second Schrödinger potential has a uniformly bounded Hessian in the terminal variable, which controls the conditional relative entropy of the optimal coupling; (ii) a regime-aware transport cost W_ω that uses Euclidean distance inside a regime and unit cost between regimes, used to measure stability of the bridge under terminal perturbations; and (iii) a recursive inequality that combines a stabili

What would settle it

A concrete test: take a regime-switching diffusion with smooth, state-dependent drift and switching rates, compactly supported marginals satisfying the Talagrand condition, and run the Sinkhorn iteration; if the relative entropy to the optimal plan does not decay geometrically—or if the transition density r_ij(0,x;T,y) is found to vanish or lose second-order smoothness on the supports—Theorem 2.12's conclusion would fail.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the Schrödinger bridge for a regime-switching reference process can be solved by Sinkhorn iteration with a guaranteed exponential rate in relative entropy. Writing the cost as c_ij(x,y) = -log r_ij(0,x;T,y) for the transition density of the reference diffusion with switching, the algorithm alternates updates of the two Schrödinger potentials. Theorem 2.12 states that if the marginal measures are compactly supported, absolutely continuous, and the terminal one satisfies a Talagrand transport inequality, and if the transition density is C^2 and strictly positive on the relevant supports, then there exist C>0 and θ∈(0,1) such that H(π^{ρ,μ}|π^{n+1

Load-bearing premise

The main theorems rest on Assumption 2.5—that the reference process's transition density is C^2 and strictly positive on the product of the supports of the prescribed marginals—and the paper verifies this regularity only for constant-coefficient dynamics, leaving the general state-dependent case as an unproven premise.

Editorial extensions

If this is right

  • The Sinkhorn algorithm for regime-switching Schrödinger bridges is guaranteed to converge geometrically in relative entropy, so in practice the iteration can be stopped after a predictable number of steps under the stated assumptions.
  • The result covers the partially observed terminal setting, where only the continuous component is prescribed; after the identified state-space extension, the same exponential rate applies.
  • The stability estimate Theorem 2.10 yields quantitative control of the optimal bridge when the terminal distribution is perturbed, with constants depending on cost bounds and the Talagrand constant.
  • Because the relative-entropy distances are preserved when the endpoint couplings are lifted to path measures, the exponential convergence transfers to the reconstructed Schrödinger bridge on path space.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The main open gap is Assumption 2.5: the C^2 strict positivity of the transition density is verified only for constant-coefficient dynamics. For state-dependent drifts, volatilities, or switching rates, one would need a separate regularity proof (for instance via parabolic Hörmander conditions) before the exponential-rate conclusion is unconditional.
  • The paper's stability estimate suggests a practical diagnostic: monitoring the relative entropy between successive Sinkhorn iterates should show a predictable geometric pattern whenever the true bridge satisfies the assumptions; a significant deviation could indicate that the transition density regularity breaks down.
  • The partially observed setting with a dummy terminal regime effectively shows that any terminal observation that is a deterministic function of the hybrid state can be absorbed into an embedding argument; this may extend to other partial-information patterns, like observing only the regime at intermediate times.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper studies the Sinkhorn algorithm for the Schrödinger bridge problem with regime switching, on the hybrid state space E = R^d × I. The reference process is a regime-switching diffusion with state-dependent drift, volatility, and switching rates. The main results are: (i) a stability estimate in relative entropy for optimal entropic plans under perturbations of one marginal (Theorem 2.10), and (ii) exponential convergence in relative entropy of the Sinkhorn iterates, both for the fully observed terminal condition (Theorem 2.12) and for the partially observed setting where only the continuous component of the terminal marginal is prescribed (Theorems 2.20 and 2.21). The proofs follow the semiconcavity approach of Chiarini–Conforti–Greco–Tamanini and use the exponential total-variation convergence of Eckstein; several entropy identities are imported from Nutz's lecture notes [27]. The paper also contains an appendix proving C^2 strict positivity of the transition density in the constant-coefficient case.

Significance. If the main theorems are correct, this is the first quantitative convergence result for the Sinkhorn algorithm in the regime-switching Schrödinger bridge setting, extending the recent semiconcavity-based analysis from the Euclidean case to hybrid state spaces. The stability estimate in Theorem 2.10 is of independent interest, and the partially observed extension is new. The paper is clearly written and the overall proof architecture is coherent: the semiconcavity lemmas (Lemmas 3.3–3.5) are derived from first principles, the recursion in the proof of Theorem 2.12 is explicit, and the partial-observation embedding in Section 4 is a natural and elegant reduction. The main reservation is that the advertised scope is broader than what is actually proved, because Assumption 2.5 is only verified for constant coefficients; the main theorems are therefore conditional on an unproven regularity hypothesis in the general state-dependent case.

major comments (3)
  1. [§2.3, Assumption 2.5; Eq. (16); Lemma 3.4; Theorem 2.12] Assumption 2.5 — C^2 strict positivity of the transition density r_ij(0,x;T,y) on the relevant supports — is load-bearing: it makes the constants M0–M2 finite, supports the semiconcavity bound in Lemma 3.4 (Λ = 2M2 + M1^2), and is needed for the oscillation bounds and hence for the stability estimate. The paper verifies Assumption 2.5 only for constant coefficients (Example 2.9 and Appendix A). For state-dependent b_i, σ_i, and λ_ij, no proof of C^2 strict positivity is supplied. Since the abstract and introduction describe a general class of regime-switching systems, the main theorems are conditional on a regularity hypothesis that is plausible but unproven in the advertised generality. Please either prove Assumption 2.5 under natural sufficient conditions (e.g., uniform ellipticity plus Hölder regularity) or explicitly restrict the scope of the main theorems to reference processes for
  2. [§4, Proofs of Theorems 2.20 and 2.21] The proof of Theorem 2.21 is only sketched via the embedding into the fully observed problem. It claims that the Sinkhorn iterates of the partially observed problem coincide with those of the extended problem, but the verification of Assumptions 2.4 and 2.5 for the extended data (μ_p^ext concentrated on regime 1, cost c^ext extended by 0 for j≠1) is not given. In particular, one must check that the extended cost satisfies the required C^2 strict positivity and that the Talagrand condition for μ_p^ext follows from Assumption 2.18. The proof of Theorem 2.20 similarly relies on Theorem 2.10 for the extended problem. Please provide a complete argument, or at least a detailed verification of the assumptions for the extended problem, so that the partial-observation results are not conditional on an unstated extension of the fully observed theory.
  3. [§2.1–§3, Theorem 2.2, Lemma 3.1, Prop. 6.5/6.10 of [27]] Several key identities in the proof of Theorem 2.12 are imported from unpublished lecture notes [27] (version of December 2022): the characterization theorem (Theorem 2.2), the oscillation bound for optimal potentials ([27, Lemma 4.11]), the EOT structure theorem ([27, Theorem 4.2(b)]), and the entropy monotonicity identities ([27, Propositions 6.5 and 6.10]) that seed the recursion (25). Since these notes are not peer-reviewed and may not be easily accessible to all readers, please either state the needed results in the paper (in a preliminary section or appendix) or replace them with published references. This is essential for the self-containedness and verifiability of the central convergence argument.
minor comments (4)
  1. [Assumption 2.4 and Theorems 2.10, 2.12] Assumption 2.4 and the statements of Theorems 2.10 and 2.12 write ρ, μ ∈ P(R^d), but the problem is on the product space R^d × I. The notation should be corrected to P(R^d × I) or P(E).
  2. [Section 2.3, after Eq. (16)] Remark 2.6 correctly notes that only y-derivatives are needed when the T2 condition is on the terminal marginal, but the proof of Lemma 3.4 uses the full-bound notation M1 and M2. It would help readers to clarify that M1 and M2 can be read as the y-partial bounds in the proof of Theorem 2.10.
  3. [Appendix A, after Theorem A.1] In the proof of strict positivity for i≠j, the displayed Gaussian integral has a typo in the covariance argument: it should be g_{σ_i^2 s + σ_j^2 (T−s)}(y − x − b_i s − b_j(T−s)), with the variance as a function of s. This is clear from the text but should be corrected.
  4. [Proof of Theorem 2.20] The sentence 'The proof continues to employ the state-space extension method introduced in Proof of Theorem 2.21' appears before Theorem 2.21 is proved in the text. Reordering the proofs or adding a forward reference would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; convergence chain is seeded by independent TV and entropy inputs, with only a constant-coefficient verification gap for Assumption 2.5.

full rationale

The central derivation is not circular. The Sinkhorn iteration (24) and couplings (9)-(10) define a concrete algorithm, and Theorem 2.12 proves exponential entropy convergence by combining (i) the stability bound Theorem 2.10, which is proved from the semiconcavity Lemma 3.4 and Lemmas 3.5-3.6; (ii) the external total-variation contraction result Theorem 3.7 quoted from Eckstein [10]; and (iii) entropy projection/monotonicity identities cited to Nutz's lecture notes [27]. The recursion (25) expresses a_n in terms of a_{n-1} plus an exponentially decaying TV term; it does not presuppose the desired conclusion. The constants M0-M2 and B are defined from the cost under Assumption 2.5, not fitted to the target relative-entropy values, and the final constants C and theta are explicitly allowed to depend on rho, mu, P, and the initial iterates. No load-bearing self-citation appears: the only author-overlapping references ([3], [19]) are contextual, and the main external ingredients are by other authors. The genuine weakness is that Assumption 2.5 (C^2 strict positivity of the transition density r_ij) is only verified in the constant-coefficient case (Example 2.9 and Appendix A), while for state-dependent coefficients it is imposed rather than proved. That is a coverage gap in the theorem's hypotheses, not a circular reduction of the claimed result to its own inputs.

Assumptions & free parameters 0 free parameters · 5 assumptions · 2 invented entities

The central claims rest on standard domain assumptions (compact support, absolute continuity, Talagrand T2), the strong and only-partially-verified regularity Assumption 2.5, and several external results from [10] and [27]. No free parameters are fitted; the constants M0-M2 and C_μ are defined from the input data, not tuned. The invented entities are mathematical proof devices, not new physical postulates.

assumptions (5)
  • domain assumption Assumption 2.1: H(ρ⊗μ|P_{0,T})<∞ and P_{0,T}|_{supp(ρ⊗μ)} ≪ ρ⊗μ.
    Ensures the static entropic optimal transport problem is well-posed and the Schrödinger system has a solution; standard in EOT.
  • domain assumption Assumption 2.4: Conditional laws a.c., compact support, and regime-wise Talagrand T2 inequality for the terminal marginal μ with constant C_μ.
    Needed for the W_ω bound (Lemma 3.6) and for finiteness of the constants M0-M2. Restricts to compactly supported marginals.
  • domain assumption Assumption 2.5: Transition density r_{ij}(0,x;T,y) is C^2 and strictly positive on supp(ρ̂_i)×supp(μ̂_j).
    Load-bearing: underpins semiconcavity (Lemma 3.4), bounded cost constants (16), and oscillation bounds. Only verified for constant-coefficient processes in Appendix A.
  • standard math Eckstein [10, Theorem 1.2]: exponential convergence of Sinkhorn iterates in total variation under growth and tail conditions.
    Used in Corollaries 3.8-3.9 to seed the recursion in Theorem 2.12 with an exponential TV decay rate κ.
  • standard math Nutz's lecture notes [27]: Schrödinger system characterization, oscillation bounds (Lemma 4.11), EOT structure theorem (Theorem 4.2(b)), and Sinkhorn entropy identities (Propositions 6.5, 6.10).
    Background results relied on throughout; not derived in the paper. They are external, not self-citations.
invented entities (2)
  • Omega-cost W_ω on (R^d×I)^2: ω((x,i),(y,j)) = |x−y|² 1_{i=j} + 1_{i≠j}
    purpose: Controls the stability estimate with a cost that charges Euclidean squared distance within a regime and unit cost across regimes; via Lemma 3.6 it yields the final theorem.
    A proof device introduced in Section 3; well-defined mathematically, not a physical entity requiring external evidence.
  • Extended terminal marginal μ_p^ext concentrated on a dummy regime 1
    purpose: Reduces the partially observed EOT problem to a fully observed EOT problem in Theorem 2.21.
    A construction for the proof; carries no physical meaning and has no falsifiable handle outside the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Exponential Convergence of the Sinkhorn Algorithm for the Schr\"odinger Bridge with Regime Switching." pith.science (2026). https://pith.science/paper/PTPK6UIF

@misc{pith2026260719176,
  author       = {Pith},
  title        = {Pith review of: Exponential Convergence of the Sinkhorn Algorithm for the Schr\"odinger Bridge with Regime Switching},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PTPK6UIF}},
  note         = {Machine review of arXiv:2607.19176}
}
abstract

This paper studies the convergence of the Sinkhorn algorithm for the Schr\"odinger bridge problem with regime switching, as introduced in Zlotchevski and Chen (2025). We consider a class of regime-switching stochastic systems on the hybrid state space $E=\mathbb{R}^d\times\{1,\ldots,m\}$, and construct the Sinkhorn iteration through the associated equivalent entropic optimal transport formulation. The main result of this paper is the exponential convergence of the Sinkhorn algorithm in relative entropy under compactness assumptions. Our proofs are inspired by the arguments recently developed for proving exponential convergence of the classical Schr\"odinger problem in Chiarini, Conforti, Greco and Tamanini (2024), Eckstein (2025). We perform a similar analysis for the partially observed terminal setting, where only the marginal distribution of the continuous component is prescribed at the terminal time, while the discrete regime is unobserved.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 3 linked inside Pith

  1. [27]

    Introduction to entropic optimal transport, 2022

    Marcel Nutz. Introduction to entropic optimal transport, 2022. Lecture notes, version of December 5, 2022

    2022

  2. [5]

    A semiconcavity approach to stability of entropic plans and exponential convergence of Sinkhorn’s algorithm

    Alberto Chiarini, Giovanni Conforti, Giacomo Greco, and Luca Tamanini. A semiconcavity approach to stability of entropic plans and exponential convergence of Sinkhorn’s algorithm. arXiv preprint arXiv:2412.09235, 2024

    arXiv 2024

  3. [10]

    Hilbert’s projective metric for functions of bounded growth and exponential conver- gence of Sinkhorn’s algorithm.Probability Theory and Related Fields, 193(1–2):585–621, 2025

    Stephan Eckstein. Hilbert’s projective metric for functions of bounded growth and exponential conver- gence of Sinkhorn’s algorithm.Probability Theory and Related Fields, 193(1–2):585–621, 2025

    2025

  4. [1]

    Springer, Cham, 2013

    Dominique Bakry, Ivan Gentil, and Michel Ledoux.Analysis and Geometry of Markov Diffusion Oper- ators, volume 348 ofGrundlehren der mathematischen Wissenschaften. Springer, Cham, 2013

    2013

  5. [2]

    Bobkov and Michel Ledoux

    Sergey G. Bobkov and Michel Ledoux. From Brunn–Minkowski to Brascamp–Lieb and to logarithmic Sobolev inequalities.Geometric and Functional Analysis, 10(5):1028–1052, 2000

    2000

  6. [3]

    Propagation of weak log-concavity along generalised heat flows via Hamilton–Jacobi equations

    Louis-Pierre Chaintron, Giovanni Conforti, and Katharina Eichinger. Propagation of weak log-concavity along generalised heat flows via Hamilton–Jacobi equations. arXiv preprint arXiv:2508.07931, 2025

    arXiv 2025

  7. [4]

    Theodorou

    Tianrong Chen, Guan-Horng Liu, and Evangelos A. Theodorou. Likelihood training of Schr¨ odinger bridge using forward-backward SDEs theory. InInternational Conference on Learning Representations, 2022

    2022

  8. [6]

    Weak semiconvexity estimates for Schr¨ odinger potentials and logarithmic Sobolev inequality for Schr¨ odinger bridges.Probability Theory and Related Fields, 189:1045–1071, 2024

    Giovanni Conforti. Weak semiconvexity estimates for Schr¨ odinger potentials and logarithmic Sobolev inequality for Schr¨ odinger bridges.Probability Theory and Related Fields, 189:1045–1071, 2024

    2024

Show all 43 references
  1. [7]

    Quantitative contraction rates for Sinkhorn algorithm: beyond bounded costs and compact marginals

    Giovanni Conforti, Alain Oliviero Durmus, and Giacomo Greco. Quantitative contraction rates for Sinkhorn algorithm: beyond bounded costs and compact marginals. arXiv preprint arXiv:2304.04451, 2023

    2023 arXiv

  2. [8]

    Sinkhorn distances: Lightspeed computation of optimal transport.Advances in neural information processing systems, 26, 2013

    Marco Cuturi. Sinkhorn distances: Lightspeed computation of optimal transport.Advances in neural information processing systems, 26, 2013

    2013

  3. [9]

    Diffusion Schr¨ odinger bridge with applications to score-based generative modeling

    Valentin De Bortoli, James Thornton, Jeremy Heng, and Arnaud Doucet. Diffusion Schr¨ odinger bridge with applications to score-based generative modeling. InAdvances in Neural Information Processing Systems, volume 34, pages 17695–17709, 2021

    2021

  4. [11]

    Quantitative stability of regularized optimal transport and conver- gence of Sinkhorn’s algorithm.SIAM Journal on Mathematical Analysis, 54(6):5922–5948, 2022

    Stephan Eckstein and Marcel Nutz. Quantitative stability of regularized optimal transport and conver- gence of Sinkhorn’s algorithm.SIAM Journal on Mathematical Analysis, 54(6):5922–5948, 2022

    2022

  5. [12]

    Georgiou

    Asmaa Eldesoukey, Olga Movilla Miangolarra, and Tryphon T. Georgiou. An excursion onto Schr¨ odinger’s bridges: Stochastic flows with spatio-temporal marginals.IEEE Control Systems Let- ters, 8:1138–1143, 2024

    2024

  6. [13]

    Georgiou

    Asmaa Eldesoukey, Olga Movilla Miangolarra, and Tryphon T. Georgiou. Inferring evanescent stochastic dynamics from marginal distributions.Journal of Systems Science and Complexity, 38(1):129–149, 2025

    2025

  7. [14]

    Random fields and diffusion processes

    Hans F¨ ollmer. Random fields and diffusion processes. In ´Ecole d’ ´Et´ e de Probabilit´ es de Saint-Flour XV–XVII, 1985–87, volume 1362 ofLecture Notes in Mathematics, pages 101–203. Springer, Berlin, 1988

    1985

  8. [15]

    On the convergence rate of Sinkhorn’s algorithm.Mathematics of Operations Research, 51(2):1080–1096, 2025

    Promit Ghosal and Marcel Nutz. On the convergence rate of Sinkhorn’s algorithm.Mathematics of Operations Research, 51(2):1080–1096, 2025

    2025

  9. [16]

    Non-asymptotic con- vergence bounds for Sinkhorn iterates and their gradients: a coupling approach

    Giacomo Greco, Maxence Noble, Giovanni Conforti, and Alain Oliviero-Durmus. Non-asymptotic con- vergence bounds for Sinkhorn iterates and their gradients: a coupling approach. InProceedings of the Thirty Sixth Conference on Learning Theory, volume 195 ofProceedings of Machine ...

    2023

  10. [17]

    Hessian stability and convergence rates for entropic and Sinkhorn potentials via semiconcavity

    Giacomo Greco and Luca Tamanini. Hessian stability and convergence rates for entropic and Sinkhorn potentials via semiconcavity. arXiv preprint arXiv:2504.11133, 2025. 34

    2025

  11. [18]

    Risk management of guaranteed minimum benefits under a regime-switching jump-diffusion model.Numerical Algebra, Control and Optimization, 15(1):52–76, 2025

    Wenlong Hu and Tao Pang. Risk management of guaranteed minimum benefits under a regime-switching jump-diffusion model.Numerical Algebra, Control and Optimization, 15(1):52–76, 2025

    2025

  12. [19]

    Entropic optimal transport problem with convex functional cost

    Anna Kazeykina, Zhenjie Ren, Xiaozhen Wang, and Yufei Zhang. Entropic optimal transport problem with convex functional cost. arXiv preprint arXiv:2503.11843, 2025

    2025

  13. [20]

    N. V. Krylov.Lectures on Elliptic and Parabolic Equations in H¨ older Spaces, volume 12 ofGraduate Studies in Mathematics. American Mathematical Society, Providence, RI, 1996

    1996

  14. [21]

    O. A. Ladyzhenskaya, V. A. Solonnikov, and N. N. Ural’tseva.Linear and Quasi-linear Equations of Parabolic Type, volume 23 ofTranslations of Mathematical Monographs. American Mathematical Society, Providence, Rhode Island, 1968

    1968

  15. [22]

    A survey of the Schr¨ odinger problem and some of its connections with optimal transport.Discrete and Continuous Dynamical Systems - A, 34(4):1533–1574, 2014

    Christian L´ eonard. A survey of the Schr¨ odinger problem and some of its connections with optimal transport.Discrete and Continuous Dynamical Systems - A, 34(4):1533–1574, 2014

    2014

  16. [23]

    Lieberman.Second Order Parabolic Differential Equations

    Gary M. Lieberman.Second Order Parabolic Differential Equations. World Scientific, 1996

    1996

  17. [24]

    I 2SB: Image-to-image Schr¨ odinger bridge

    Guan-Horng Liu, Arash Vahdat, De-An Huang, Evangelos Theodorou, Weili Nie, and Anima Anand- kumar. I 2SB: Image-to-image Schr¨ odinger bridge. In Andreas Krause, Emma Brunskill, Kyunghyun Cho, Barbara Engelhardt, Sivan Sabato, and Jonathan Scarlett, editors,Proceedings of the ...

    2023

  18. [25]

    Yu-Ying Liu, Shuang Li, Fuxin Li, Le Song, and James M. Rehg. Efficient learning of continuous-time hidden Markov models for disease progression. InAdvances in Neural Information Processing Systems, 2015

    2015

  19. [26]

    Stochastic population dynamics under regime switching.Journal of Mathe- matical Analysis and Applications, 334(1):69–84, October 2007

    Qi Luo and Xuerong Mao. Stochastic population dynamics under regime switching.Journal of Mathe- matical Analysis and Applications, 334(1):69–84, October 2007

    2007

  20. [28]

    Stability of Schr¨ odinger potentials and convergence of Sinkhorn’s algorithm.The Annals of Probability, 51(2):699–722, 2023

    Marcel Nutz and Johannes Wiesel. Stability of Schr¨ odinger potentials and convergence of Sinkhorn’s algorithm.The Annals of Probability, 51(2):699–722, 2023

    2023

  21. [29]

    Oxtoby et al

    Neil P. Oxtoby et al. Data-driven disease progression modeling. InMachine Learning for Brain Disor- ders. NCBI Bookshelf, 2023

    2023

  22. [30]

    Erwin Schr¨ odinger.¨Uber die umkehrung der naturgesetze.Sitzungsberichte der Preussischen Akademie der Wissenschaften, pages 144–153, 1931

    1931

  23. [31]

    Sur la th´ eorie relativiste de l’´ electron et l’interpr´ etation de la m´ ecanique quantique

    Erwin Schr¨ odinger. Sur la th´ eorie relativiste de l’´ electron et l’interpr´ etation de la m´ ecanique quantique. Ann. Inst. H. Poincar´ e, 2(4):269–310, 1932

    1932

  24. [32]

    Aligned diffusion Schr¨ odinger bridges

    Vignesh Ram Somnath, Matteo Pariset, Ya-Ping Hsieh, Maria Rodriguez Martinez, Andreas Krause, and Charlotte Bunne. Aligned diffusion Schr¨ odinger bridges. In Robin J. Evans and Ilya Shpitser, editors,Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Inte...

    1985

  25. [33]

    Tay et al

    E. Tay et al. Shape modeling of longitudinal medical images: from diffeomorphic metric mapping to deep learning.Frontiers in Artificial Intelligence, 2025

    2025

  26. [34]

    Therneau and Fang-Shu Ou

    Terry M. Therneau and Fang-Shu Ou. Using multistate models with clinical trial data for a deeper understanding of complex disease processes.Clinical Trials, 2024

    2024

  27. [35]

    Tong, Nikolay Malkin, Kilian Fatras, Lazar Atanackovic, Yanlei Zhang, Guillaume Huguet, Guy Wolf, and Yoshua Bengio

    Alexander Y. Tong, Nikolay Malkin, Kilian Fatras, Lazar Atanackovic, Yanlei Zhang, Guillaume Huguet, Guy Wolf, and Yoshua Bengio. Simulation-free Schr¨ odinger bridges via score and flow matching. In Sanjoy Dasgupta, Stephan Mandt, and Yingzhen Li, editors,Proceedings of The 2...

    2024

  28. [36]

    T. D. Tuong, Dang H. Nguyen, N. T. Dieu, and Ky Tran. Extinction and permanence in a stochastic SIRS model in regime-switching with general incidence rate.Nonlinear Analysis: Hybrid Systems, 34:121–130, 2019

    2019

  29. [37]

    Department of Health and Human Services

    U.S. Department of Health and Human Services. Telehealth research recap: Remote patient monitoring, September 2024. Telehealth.HHS.gov research recap

    2024

  30. [38]

    Deep generative learning via Schr¨ odinger bridge

    Gefei Wang, Yuling Jiao, Qian Xu, Yang Wang, and Can Yang. Deep generative learning via Schr¨ odinger bridge. In Marina Meila and Tong Zhang, editors,Proceedings of the 38th International Conference on Machine Learning, volume 139 ofProceedings of Machine Learning Research, pa...

    2021

  31. [39]

    George Yin and Chao Zhu.Hybrid Switching Diffusions: Properties and Applications, volume 63 of Stochastic Modelling and Applied Probability

    G. George Yin and Chao Zhu.Hybrid Switching Diffusions: Properties and Applications, volume 63 of Stochastic Modelling and Applied Probability. Springer, New York, 2010

    2010

  32. [40]

    Stock trading: An optimal selling rule.SIAM Journal on Control and Optimization, 40(1):64–87, 2001

    Qing Zhang. Stock trading: An optimal selling rule.SIAM Journal on Control and Optimization, 40(1):64–87, 2001

    2001

  33. [41]

    The Schr¨ odinger bridge problem for jump diffusions

    Andrei Zlotchevski and Linan Chen. The Schr¨ odinger bridge problem for jump diffusions. arXiv preprint arXiv:2411.13765, 2024

    2024 arXiv

  34. [42]

    A regime-switching approach to the unbalanced Schr¨ odinger bridge problem

    Andrei Zlotchevski and Linan Chen. A regime-switching approach to the unbalanced Schr¨ odinger bridge problem. arXiv preprint arXiv:2512.12971, 2025

    2025

  35. [43]

    The Schr¨ odinger bridge problem for jump diffusions with regime switching

    Andrei Zlotchevski and Linan Chen. The Schr¨ odinger bridge problem for jump diffusions with regime switching. arXiv preprint arXiv:2511.06079, 2025. 36

    2025

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.