Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Categorical Schr\"odinger Bridge Matching

T0 review · 3 major / 6 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read In finite discrete spaces with full-support marginals and a Markov reference, the dynamic Schrödinger bridge is the unique process that is both Markovian and reciprocal, and the D-IMF alternating-projection scheme provably converges to it…

desk verdict A genuinely new characterization theorem for Schrödinger bridges on discrete spaces, paired with an algorithm whose convergence guarantee is cited rather than proved. read the letter →

arxiv 2502.01416 v4 pith:RRWOBAVJ submitted 2025-02-03 cs.LG

Grigoriy Ksenofontov , Alexander Korotin This is my paper · ORCID
classification cs.LG
keywords SchrödingerbridgediscretestatespacesiterativeMarkovianfittingunpaireddomaintranslationcategoricalgenerativemodelingvectorquantizationentropicoptimaltransport
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 establishes that Schrödinger bridge problems on finite discrete state spaces are solvable by a specific alternating-projection scheme with a convergence guarantee. Its main theorem shows that, for full-support marginals and a full-support Markov reference process, the unique solution of the dynamic Schrödinger bridge is the only process that is simultaneously Markovian and reciprocal with respect to the reference. A corollary gives that discrete-time Iterative Markovian Fitting (D-IMF), which alternates reciprocal and Markovian projections, converges to this solution in Kullback–Leibler divergence. The authors then turn this guarantee into a practical algorithm, Categorical Schrödinger Bridge Matching (CSBM), and demonstrate it on synthetic data, vector-quantized images, and text sentiment transfer.

What carries the argument

The central object is a finite-state discrete-time process that is simultaneously Markovian and reciprocal. Reciprocity means that, conditionally on the two endpoints, the interior path follows the same bridge as the reference process, while Markovianity means transitions factor one step at a time. The load-bearing identity is equation (11): for any intermediate time $t_n$, $\log q(x_1|x_0)-\log q_{\mathrm{ref}}(x_0,x_1)$ splits as $f_0(x_0,x_{t_n})+f_1(x_{t_n},x_1)$, which forces the endpoint coupling to have the Schrödinger-bridge form $\psi(x_0)\,q_{\mathrm{ref}}(x_1|x_0)\,\phi(x_1)$. The algorithmic machinery is D-IMF, the alternating reciprocal and Markovian projections, implemented with a neural network that samples an endpoint from a factorized learned distribution and then uses the reference bridge to move between states.

What would settle it

Run D-IMF analytically on a small finite state space with a full-support Markov reference and compute the exact Sinkhorn bridge $q^*$. If the iterates' $\mathrm{KL}(q^l\,\|\,q^*)$ does not approach zero, or if two different processes are simultaneously Markovian and reciprocal with the same endpoint marginals, Theorem 3.1 and Corollary 3.2 would be refuted.

Watch

Extended reading notes

Core claim

The central discovery is a characterization theorem for the dynamic Schrödinger bridge in finite discrete spaces with a general Markov reference process. Theorem 3.1 says: if $q^*$ has the prescribed endpoint marginals $p_0$ and $p_1$, is Markovian, and is reciprocal with respect to $q_{\mathrm{ref}}$ (meaning its bridges between endpoints coincide with $q_{\mathrm{ref}}$'s), then $q^*$ is the unique minimizer of $\mathrm{KL}(q\,\|\,q_{\mathrm{ref}})$ over all processes with those endpoint marginals. The proof rewrites the joint density of a Markov-reciprocal process as $q(x_0,x_1)=\psi(x_0)\,q_{\mathrm{ref}}(x_1|x_0)\,\phi(x_1)$, the canonical form of the static entropic optimal transport plan, and then invokes the standard identification of that plan as the Schrödinger bridge. Corollary 3.2 then states that D-IMF iterates converge in KL to $q^*$, which is the theoretical foundation the D-IMF procedure previously lacked for discrete spaces.

Load-bearing premise

The claim leans on the assumption that the D-IMF alternating projection scheme converges in KL to the unique Markovian-reciprocal process in finite discrete spaces, a transfer from a generic existing proof rather than a proof worked out in this paper.

Editorial extensions

If this is right

  • D-IMF now has a convergence guarantee on finite discrete spaces with a general Markov reference, not just continuous Euclidean setups.
  • CSBM can be applied to discrete data such as vector-quantized codebooks, text tokens, and categorical variables, as demonstrated on colored MNIST, CelebA latent spaces, and Amazon reviews.
  • The theory requires only a finite number of time steps, and the paper notes that even $N=1$ intermediate step is enough for the characterization to hold.
  • The practical algorithm works with both uniform and Gaussian-like categorical reference processes and inherits the diffusion-style training objective from discrete diffusion models.
  • The paper's experiments indicate that CSBM attains competitive generative quality on unpaired image translation while operating directly on discrete tokens rather than continuous pixel space.

Reading between the lines

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

  • An inference beyond the paper: the proof argument is stated to apply to general state spaces, which suggests the same characterization could yield a discrete-free variant of D-IMF for continuous spaces with arbitrary Markov references, not just the Wiener-process case.
  • An inference beyond the paper: the full-support assumption on $p_0$, $p_1$, and $q_{\mathrm{ref}}$ is load-bearing, because the proof takes logarithms of strictly positive densities; extending the result to sparse or degenerate marginals may require a limiting argument or an explicit support-conditioned formulation.
  • An inference beyond the paper: the algorithmic factorization over dimensions is not covered by the theorem, so the visual pixelation reported in the paper likely reflects this implementation gap rather than the convergence result; a testable extension would compare factorized transitions against copula-based or energy-based joint transitions on the same tasks.
  • An inference beyond the paper: the paper observes that convergence behavior depends on the number of time steps $N$, so a quantitative convergence-rate analysis as a function of $N$, $\alpha$, and the reference process would be a natural next step.
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 / 6 minor

Summary. The paper studies the Schrödinger Bridge (SB) problem on finite discrete spaces X = S^D with a general full-support Markov reference process. Its main theoretical contribution is Theorem 3.1, which states that in this setting the dynamic SB solution q* is the unique process that is both Markov and reciprocal with respect to the reference. Based on this characterization, Corollary 3.2 claims that the discrete-time Iterative Markovian Fitting (D-IMF) procedure converges to q* in KL divergence. The paper then proposes a practical algorithm, Categorical Schrödinger Bridge Matching (CSBM), which implements the reciprocal and Markov projections using neural networks with a factorized parameterization, and reports experiments on synthetic data, colored MNIST, CelebA (via VQ-GAN latent codebooks), and Amazon Reviews sentiment transfer. The paper includes code and extensive appendices.

Significance. If the main claims hold, the paper fills a genuine gap in the Schrödinger Bridge literature: the discrete-state, discrete-time setting with a general Markov reference was listed as an open case in the authors' Table 1, whereas continuous-space analogues were covered by prior work (Shi et al. 2023; Gushchin et al. 2024b). The characterization theorem is clean and its proof is mostly self-contained, citing Leonard's standard result for the static SB. The proposed CSBM algorithm is a natural extension of D3PM-style discrete diffusion to a Schrödinger bridge setting, and the authors provide code and a range of experiments. However, the algorithmic convergence guarantee—the paper's central practical claim—is not actually proved, and the experimental comparison against continuous-space baselines is confounded by operating in different data spaces. These issues are load-bearing for the paper's core narrative, but they are fixable within the manuscript's scope.

major comments (3)
  1. [Section 4.4, Table 2] The convergence of D-IMF is asserted with the sentence 'the convergence easily follows from the generic proof argument in (Shi et al., 2023, Theorem 8)' and no proof is provided. Shi et al.'s Theorem 8 is a continuous-time result, and its proof relies on continuous-time machinery (Brownian bridges, Girsanov-type arguments) that does not automatically transfer to the finite discrete state space and arbitrary full-support Markov reference qref considered here. Theorem 3.1 characterizes q* but does not by itself establish that the alternating KL projections in Eq. (6) converge, or that the limit is q* rather than some other fixed point. Since Corollary 3.2 is the paper's core algorithmic guarantee, this is not a local omission: the authors should either give a self-contained proof for finite spaces (e.g., by viewing Eq. (6) as cyclic KL projections and invoking a finite-state alternating minimization theorem, or by adapting the argument of Shi et al. explicitly to this setting) or state a precise theorem with conditions under which convergence is guaranteed and defer to a complete proof.
  2. [Section 4.1] The quantitative comparison with ASBM and DSBM is not apples-to-apples: CSBM operates on discrete VQ-GAN latent tokens (S=1024, D=256), whereas ASBM and DSBM operate in continuous pixel space, and the comparison numbers are taken from (Gushchin et al., 2024b) rather than re-evaluated in a shared protocol. The better FID/CMMD/LPIPS values for CSBM could partly reflect the information already lost or organized in the VQ-GAN latent space, not a superiority of the SB solver. In addition, the CSBM metrics in Table 2 are reported without error bars or multiple seeds, so the claim that 'our approach achieves better results' is not statistically substantiated. I recommend reframing these results as an illustration rather than a head-to-head win, or providing a same-space comparison (e.g., decoding both methods' outputs to pixels and then computing metrics on the same decoded images).
  3. [Section 4.1] The numerical verification of D-IMF convergence is restricted to the symmetric qunif and qgauss reference processes. Corollary 3.2 claims convergence for any full-support Markov qref. Please add at least one asymmetric, non-reversible reference process (e.g., a Markov chain with a drift) to the synthetic convergence study, or explicitly state that the general case is not empirically tested. Without such a test, the paper's evidence for the broad claim in Corollary 3.2 is incomplete.
minor comments (6)
  1. [Figure 1 caption] The word 'stochastisity' should be 'stochasticity'.
  2. [Section 4, first paragraph] 'additional immages' should be 'additional images'.
  3. [Eq. (10)] The displayed L(m) has a minus sign before the term E_{qref(x_{t_N}|x_0,x_1)}[log m(x_1|x_{t_N})] that follows from the KL decomposition; the formatting makes it look like a separate term, so please clarify the derivation or restructure the equation.
  4. [Algorithm 1] In the forward step, sampling n ~ U[1, N+1] and then sampling x_{t_{n-1}} ~ qref(x_{t_{n-1}}|x_0,x_1) is ambiguous when n = N+1 because x_{t_N} is not defined as a state before x_1 in the forward chain; please clarify the indexing (e.g., define x_{t_0}=x_0, x_{t_{N+1}}=x_1, or restrict n to 1..N).
  5. [Table 5] The rows 'D-IMF=1 grad updates' and 'D-IMF grad updates' are confusing; consider renaming them 'gradient updates per D-IMF sweep' and 'total gradient updates per outer iteration'.
  6. [Appendix C.4] The statement 'manual inspection of the samples in Table 4 suggests that most generations convey the correct polarity' is subjective; reporting a quantitative polarity score on the generated samples would be more convincing.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the main characterization is derived from first principles against external static-SB theory; the D-IMF convergence corollary is an unproved transfer from Shi et al., not a circular reduction.

full rationale

The derivation chain is not circular. Theorem 3.1 starts from the assumed Markov and reciprocal properties of q* and derives log q(x1|x0) - log qref(x0,x1) = f0(x0,xtn) + f1(xtn,x1); fixing x-dagger shows the additive separation into g0(x0)+g1(x1), yielding q(x0,x1) = psi(x0) qref(x1|x0) phi(x1). The identification of this factorized form with the static Schrodinger Bridge solution is imported from Leonard (2013, Theorem 2.8), an external measure-theoretic result, not from the present authors' prior work. The dynamic conclusion then follows because q* carries the reference conditional bridge, i.e. is reciprocal, which is the standard decomposition in Eq. (4), not an assumption equivalent to the conclusion. Corollary 3.2 is the one genuinely unsupported step: it asserts KL convergence of the D-IMF iterates by saying it 'easily follows from the generic proof argument in (Shi et al., 2023, Theorem 8)' without supplying the transfer argument for finite state spaces and general Markov qref. That is an omitted proof and a transfer risk, but it is not circular: Shi et al. is external, and the convergence claim is not obtained by renaming a fitted parameter or by defining q* as the limit of Eq. (6). Self-citations to Gushchin et al. (2024b) define the D-IMF projections and supply Proposition 3.5 used in Proposition 3.3; these are non-load-bearing for the main characterization, and the cited claims are independently checkable mathematical statements. The free parameters alpha and lambda are tuning choices, and the convergence experiment uses the Sinkhorn algorithm as an external ground truth. Appendix A itself flags the factorization limitation and a negligible implementation discrepancy; those are limitations, not circular reductions. No step reduces by construction to its own input, so the paper receives a low score reflecting only minor, non-load-bearing self-citations.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

No new physical or mathematical entities are introduced. The reference processes qunif/qgauss are standard constructions. The free parameters are experimental design choices.

free parameters (3)
  • stochasticity parameter α = per-experiment: 0.005, 0.01 for CelebA; 0.01, 0.05 for MNIST; etc.
    Controls the reference process transition probabilities in qunif/qgauss and is tuned per dataset (Appendix D.1). Directly affects the learned bridge and final quality.
  • loss weight λ (L_simple) = 0.001 for all experiments
    Weight for the auxiliary simple loss in Eq. (21) and (22), chosen by hand (Appendix D.3).
  • number of time steps N = varies: 2,4,10,25,50,100
    Number of intermediate time moments, chosen per experiment; affects reference process and model capacity.
assumptions (3)
  • domain assumption The reference process qref is Markov with full support on X^{N+2}, and p0, p1 have full support.
    Stated in Theorem 3.1; needed for the Bayes factorization qref(xtn|x0,x1)=qref(x0)qref(xtn|x0)qref(x1|xtn)/qref(x0,x1) and positive densities.
  • standard math The static Schrödinger Bridge solution has the form q(x0,x1)=ψ(x0)qref(x1|x0)φ(x1) (Leonard 2013, Theorem 2.8).
    Used in the proof of Theorem 3.1 to conclude that the constructed process is the SB solution.
  • domain assumption The D-IMF alternating projection scheme (Eq. 6) converges in KL to the unique Markovian-reciprocal process; this is assumed to transfer from Shi et al. 2023, Theorem 8.
    Corollary 3.2 relies on this without a proof specific to finite discrete spaces; if the convergence fails, the algorithm may not reach the SB.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Categorical Schr\"odinger Bridge Matching." pith.science (2026). https://pith.science/paper/RRWOBAVJ

@misc{pith2026250201416,
  author       = {Pith},
  title        = {Pith review of: Categorical Schr\"odinger Bridge Matching},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RRWOBAVJ}},
  note         = {Machine review of arXiv:2502.01416}
}
abstract

The Schr\"odinger Bridge (SB) is a powerful framework for solving generative modeling tasks such as unpaired domain translation. Most SB-related research focuses on continuous data space $\mathbb{R}^{D}$ and leaves open theoretical and algorithmic questions about applying SB methods to discrete data, e.g, on finite spaces $\mathbb{S}^{D}$. Notable examples of such sets $\mathbb{S}$ are codebooks of vector-quantized (VQ) representations of modern autoencoders, tokens in texts, categories of atoms in molecules, etc. In this paper, we provide a theoretical and algorithmic foundation for solving SB in discrete spaces using the recently introduced Iterative Markovian Fitting (IMF) procedure. Specifically, we theoretically justify the convergence of discrete-time IMF (D-IMF) to SB in discrete spaces. This enables us to develop a practical computational algorithm for SB, which we call Categorical Schr\"odinger Bridge Matching (CSBM). We show the performance of CSBM via a series of experiments with synthetic data and VQ representations of images. The code of CSBM is available at https://github.com/gregkseno/csbm.

Figures

Figures reproduced from arXiv: 2502.01416 by the authors.

Figure 1
Figure 1. Dependence of convergence of D-IMF procedure on dis￾crete data under different N, α and q ref . the Sinkhorn algorithm (Cuturi, 2013). As a cost matrix, we use the negative logarithm of a cumulative transition matrix QN+1 n=1 Qn. The resulting convergence curves, shown in [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 3
Figure 3. Results of colored digits unpaired translation “3” → “2” learned by our CSBM algorithm with reference process q gauss and varying number of time moments N. Specifically, each pixel is represented using three 8-bit chan￾nels (RGB), i.e., S = 256, and the data space is of size 256D, where D = 32 × 32 × 3. The goal of this experiment is to evaluate the capability of CSBM to perform unpaired translation with different n… view at source ↗
Figure 4
Figure 4. Comparison of male → female translation on the CelebA 128 × 128 dataset using CSBM (ours), ASBM, and DSBM. ASBM and DSBM operate in continuous pixel space, whereas CSBM operates in a discrete latent space of VQ-GAN (Esser et al., 2021). The low-stochasticity setting for CSBM corresponds to α = 0.005, while the high-stochasticity setting corresponds to α = 0.01 of the reference process q unif. The images for ASBM and… view at source ↗
Figures from the paper (9 more)
Figure 5
Figure 5. Figure 5: SB between 2D Gaussian and Swiss-Roll distributions learned by our CSBM algorithm with MSE loss in Equations (19) and (20) for different reference processes q unif and q gauss with varying parameters α. The Alternative Objective. Analogous reasoning extends to the alte…
Figure 6
Figure 6. Figure 6: Results of colored digits unpaired translation learned by our CSBM algorithm with reference process q unif and varying stochasticity parameter α. C.3 Continuous Methods in Latent Space For completeness, we also trained DSBM in the latent space. For a fair comparison, w…
Figure 7
Figure 7. Figure 7: Results of training DSBM (Shi et al., 2023) on VQ-GAN lantent space of CelebA. The VQ-GAN model is the same as in the main experiments (M4.4). 16 [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: Results of colored digits unpaired translation “2” → “3” learned by our CSBM algorithm with reference process q gauss and varying number of time moments N. 20 [PITH_FULL_IMAGE:figures/full_fig_p020_8.png]
Figure 9
Figure 9. Figure 9: Comparison of female → male translation on the CelebA 128 × 128 dataset using CSBM (ours), ASBM, and DSBM. The low-stochasticity setting for CSBM corresponds to α = 0.005, while the high-stochasticity setting corresponds to α = 0.01. The stochasticity parameters for AS…
Figure 10
Figure 10. Figure 10: male → female translation trajectories on the CelebA 128×128 dataset using CSBM with α = 0.01. Each column corresponds to time moments 0, 10, 25, 50, 75, 90, and 101. 22 [PITH_FULL_IMAGE:figures/full_fig_p022_10.png]
Figure 11
Figure 11. Figure 11: male → female translation trajectories on the CelebA 128 × 128 dataset using CSBM with α = 0.005. Each column corresponds to time moments 0, 10, 25, 50, 75, 90, and 101. 23 [PITH_FULL_IMAGE:figures/full_fig_p023_11.png]
Figure 12
Figure 12. Figure 12: female → male translation trajectories on the CelebA 128×128 dataset using CSBM with α = 0.01. Each column corresponds to time moments 0, 10, 25, 50, 75, 90, and 101. 24 [PITH_FULL_IMAGE:figures/full_fig_p024_12.png]
Figure 13
Figure 13. Figure 13: female → male translation trajectories on the CelebA 128 × 128 dataset using CSBM with α = 0.005. Each column corresponds to time moments 0, 10, 25, 50, 75, 90, and 101. 25 [PITH_FULL_IMAGE:figures/full_fig_p025_13.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Discrete Diffusion Models: A Unified Framework from Tokenization to Generation

    cs.LG 2026-07 unverdicted novelty 4.0 of 10

    Discrete diffusion models are re-framed as instances of a tokenization-centric, four-component design space (corruption, denoiser, objective, sampler) in a broad survey with no new experimental or theoretical results.

Reference graph

Works this paper leans on

66 extracted references · 41 canonical work pages · cited by 1 Pith paper

  1. [1]

    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 format.date year duplicate empty "emp...

  2. [2]

    D., Ho, J., Tarlow, D., and Van Den Berg, R

    Austin, J., Johnson, D. D., Ho, J., Tarlow, D., and Van Den Berg, R. Structured denoising diffusion models in discrete state-spaces. Advances in Neural Information Processing Systems, 34: 0 17981--17993, 2021

    2021

  3. [3]

    S., and Ghosh, J

    Banerjee, A., Merugu, S., Dhillon, I. S., and Ghosh, J. Clustering with Bregman divergences. Journal of machine learning research, 6 0 (Oct): 0 1705--1749, 2005

    work page 2005

  4. [4]

    Generative flows on discrete state-spaces: Enabling multimodal flows with applications to protein co-design

    Campbell, A., Yim, J., Barzilay, R., Rainforth, T., and Jaakkola, T. Generative flows on discrete state-spaces: Enabling multimodal flows with applications to protein co-design. In Salakhutdinov, R., Kolter, Z., Heller, K., Weller, A., Oliver, N., Scarlett, J., and Berkenkamp, F. (eds.), Proceedings of the 41st International Conference on Machine Learning...

    work page 2024

  5. [5]

    Likelihood training of Schr \"o dinger bridge using forward-backward SDEs theory

    Chen, T., Liu, G.-H., and Theodorou, E. Likelihood training of Schr \"o dinger bridge using forward-backward SDEs theory. In International Conference on Learning Representations, 2022

    work page 2022

  6. [6]

    Sinkhorn distances: Lightspeed computation of optimal transport

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

    2013

  7. [7]

    Diffusion Schr \"o dinger bridge with applications to score-based generative modeling

    De Bortoli, V., Thornton, J., Heng, J., and Doucet, A. Diffusion Schr \"o dinger bridge with applications to score-based generative modeling. Advances in Neural Information Processing Systems, 34: 0 17695--17709, 2021

    work page 2021

  8. [8]

    Schr\"odinger bridge flow for unpaired data translation

    De Bortoli, V., Korshunova, I., Mnih, A., and Doucet, A. Schr\"odinger bridge flow for unpaired data translation. In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. URL https://openreview.net/forum?id=1F32iCJFfa

    work page 2024

Show all 66 references
  1. [9]

    Rates of estimation of optimal transport maps using plug-in estimators via barycentric projections

    Deb, N., Ghosal, P., and Sen, B. Rates of estimation of optimal transport maps using plug-in estimators via barycentric projections. Advances in Neural Information Processing Systems, 34: 0 29736--29753, 2021

    2021

  2. [10]

    Decentralize and randomize: Faster algorithm for W asserstein barycenters

    Dvurechenskii, P., Dvinskikh, D., Gasnikov, A., Uribe, C., and Nedich, A. Decentralize and randomize: Faster algorithm for W asserstein barycenters. In Advances in Neural Information Processing Systems, pp.\ 10760--10770, 2018

    2018

  3. [11]

    Computational optimal transport: Complexity by accelerated gradient descent is better than by Sinkhorn ’s algorithm

    Dvurechensky, P., Gasnikov, A., and Kroshnin, A. Computational optimal transport: Complexity by accelerated gradient descent is better than by Sinkhorn ’s algorithm. In International conference on machine learning, pp.\ 1367--1376. PMLR, 2018

    2018

  4. [12]

    Taming transformers for high-resolution image synthesis

    Esser, P., Rombach, R., and Ommer, B. Taming transformers for high-resolution image synthesis. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 12873--12883, 2021

    2021

  5. [13]

    T., Synnaeve, G., Adi, Y., and Lipman, Y

    Gat, I., Remez, T., Shaul, N., Kreuk, F., Chen, R. T., Synnaeve, G., Adi, Y., and Lipman, Y. Discrete flow matching. In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

    2024

  6. [14]

    Generative adversarial nets

    Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., and Bengio, Y. Generative adversarial nets. In Advances in neural information processing systems, pp.\ 2672--2680, 2014

    2014

  7. [15]

    Vector quantized diffusion model for text-to-image synthesis

    Gu, S., Chen, D., Bao, J., Wen, F., Zhang, B., Chen, D., Yuan, L., and Guo, B. Vector quantized diffusion model for text-to-image synthesis. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 10696--10706, 2022

    2022

  8. [16]

    Entropic neural optimal transport via diffusion processes

    Gushchin, N., Kolesov, A., Korotin, A., Vetrov, D., and Burnaev, E. Entropic neural optimal transport via diffusion processes. In Advances in Neural Information Processing Systems, 2023 a

    2023

  9. [17]

    Building the bridge of Schr\"odinger : A continuous entropic optimal transport benchmark

    Gushchin, N., Kolesov, A., Mokrov, P., Karpikova, P., Spiridonov, A., Burnaev, E., and Korotin, A. Building the bridge of Schr\"odinger : A continuous entropic optimal transport benchmark. In Thirty-seventh Conference on Neural Information Processing Systems Datasets and Bench...

    2023

  10. [18]

    Light and optimal Schr \"o dinger bridge matching

    Gushchin, N., Kholkin, S., Burnaev, E., and Korotin, A. Light and optimal Schr \"o dinger bridge matching. In Forty-first International Conference on Machine Learning, 2024 a

    2024

  11. [19]

    Adversarial Schr\"odinger bridge matching

    Gushchin, N., Selikhanovych, D., Kholkin, S., Burnaev, E., and Korotin, A. Adversarial Schr\"odinger bridge matching. In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024 b . URL https://openreview.net/forum?id=L3Knnigicu

    2024

  12. [20]

    A probabilistic formulation of unsupervised text style transfer

    He, J., Wang, X., Neubig, G., and Berg-Kirkpatrick, T. A probabilistic formulation of unsupervised text style transfer. In International Conference on Learning Representations, 2020

    2020

  13. [21]

    GAN s trained by a two time-scale update rule converge to a local Nash equilibrium

    Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., and Hochreiter, S. GAN s trained by a two time-scale update rule converge to a local Nash equilibrium. In Advances in neural information processing systems, pp.\ 6626--6637, 2017

    2017

  14. [22]

    Denoising diffusion probabilistic models

    Ho, J., Jain, A., and Abbeel, P. Denoising diffusion probabilistic models. Advances in Neural Information Processing Systems, 33: 0 6840--6851, 2020

    2020

  15. [23]

    Argmax flows and multinomial diffusion: Learning categorical distributions

    Hoogeboom, E., Nielsen, D., Jaini, P., Forr \'e , P., and Welling, M. Argmax flows and multinomial diffusion: Learning categorical distributions. Advances in Neural Information Processing Systems, 34: 0 12454--12465, 2021

    2021

  16. [24]

    and Rigollet, P

    H \"u tter, J.-C. and Rigollet, P. Minimax estimation of smooth optimal transport maps. 2021

    2021

  17. [25]

    Categorical reparameterization with Gumbel -softmax

    Jang, E., Gu, S., and Poole, B. Categorical reparameterization with Gumbel -softmax. In International Conference on Learning Representations, 2017

    2017

  18. [26]

    Rethinking FID : Towards a better evaluation metric for image generation

    Jayasumana, S., Ramalingam, S., Veit, A., Glasner, D., Chakrabarti, A., and Kumar, S. Rethinking FID : Towards a better evaluation metric for image generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 9307--9315, 2024

    2024

  19. [27]

    Diffusion & adversarial Schr\"odinger bridges via iterative proportional Markovian fitting

    Kholkin, S., Ksenofontov, G., Li, D., Kornilov, N., Gushchin, N., Burnaev, E., and Korotin, A. Diffusion & adversarial Schr\"odinger bridges via iterative proportional Markovian fitting. arXiv preprint arXiv:2410.02601, 2024

    2024

  20. [28]

    H., Kim, S., Moon, S., Kim, H., Woo, J., and Kim, W

    Kim, J. H., Kim, S., Moon, S., Kim, H., Woo, J., and Kim, W. Y. Discrete diffusion Schr\"odinger bridge matching for graph transformation. arXiv preprint arXiv:2410.01500, 2024

    2024 arXiv

  21. [29]

    Light Schr \"o dinger bridge

    Korotin, A., Gushchin, N., and Burnaev, E. Light Schr \"o dinger bridge. In The Twelfth International Conference on Learning Representations, 2024

    2024

  22. [30]

    and Richardson, J

    Kudo, T. and Richardson, J. SentencePiece : A simple and language independent subword tokenizer and detokenizer for neural text processing. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pp.\ 66--71, 2018

    2018

  23. [31]

    A survey of the Schr\"odinger problem and some of its connections with optimal transport

    L \'e onard, C. A survey of the Schr\"odinger problem and some of its connections with optimal transport. arXiv preprint arXiv:1308.0215, 2013

    2013 arXiv

  24. [32]

    Reciprocal processes

    L \'e onard, C., R lly, S., and Zambrini, J.-C. Reciprocal processes. a measure-theoretical point of view. Probability Surveys, 11: 0 237--269, 2014

    2014

  25. [33]

    Delete, retrieve, generate: a simple approach to sentiment and style transfer

    Li, J., Jia, R., He, H., and Liang, P. Delete, retrieve, generate: a simple approach to sentiment and style transfer. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Lo...

    2018

  26. [34]

    Liu, A., Broadrick, O., Niepert, M., and Broeck, G. V. d. Discrete copula diffusion. arXiv preprint arXiv:2410.01949, 2024

    2024 arXiv

  27. [35]

    Deep generalized Schr \"o dinger bridge

    Liu, G.-H., Chen, T., So, O., and Theodorou, E. Deep generalized Schr \"o dinger bridge. Advances in Neural Information Processing Systems, 35: 0 9374--9388, 2022 a

    2022

  28. [36]

    I ^2 sb: Image-to-image Schr \"o dinger bridge

    Liu, G.-H., Vahdat, A., Huang, D.-A., Theodorou, E., Nie, W., and Anandkumar, A. I ^2 sb: Image-to-image Schr \"o dinger bridge. In International Conference on Machine Learning, pp.\ 22042--22062. PMLR, 2023

    2023

  29. [37]

    Flow straight and fast: Learning to generate and transfer data with rectified flow

    Liu, X., Gong, C., et al. Flow straight and fast: Learning to generate and transfer data with rectified flow. In The Eleventh International Conference on Learning Representations, 2022 b

    2022

  30. [38]

    Towards fine-grained text sentiment transfer

    Luo, F., Li, P., Yang, P., Zhou, J., Tan, Y., Chang, B., Sui, Z., and Sun, X. Towards fine-grained text sentiment transfer. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pp.\ 2013--2022, 2019

    2013

  31. [39]

    CrystalFlow : A flow-based generative model for crystalline materials

    Luo, X., Wang, Z., Lv, J., Wang, L., Wang, Y., and Ma, Y. CrystalFlow : A flow-based generative model for crystalline materials. arXiv preprint arXiv:2412.11693, 2024

    2024

  32. [40]

    Plugin estimation of smooth optimal transport maps

    Manole, T., Balakrishnan, S., Niles-Weed, J., and Wasserman, L. Plugin estimation of smooth optimal transport maps. The Annals of Statistics, 52 0 (3): 0 966--998, 2024

    2024

  33. [41]

    Energy-guided entropic neural optimal transport

    Mokrov, P., Korotin, A., Kolesov, A., Gushchin, N., and Burnaev, E. Energy-guided entropic neural optimal transport. In The Twelfth International Conference on Learning Representations, 2024. URL https://openreview.net/forum?id=d6tUsZeVs7

    2024

  34. [42]

    Balancing the style-content trade-off in sentiment transfer using polarity-aware denoising

    Mukherjee, S., Kasner, Z., and Du s ek, O. Balancing the style-content trade-off in sentiment transfer using polarity-aware denoising. In International Conference on Text, Speech, and Dialogue, pp.\ 172--186. Springer, 2022

    2022

  35. [43]

    Justifying recommendations using distantly-labeled reviews and fine-grained aspects

    Ni, J., Li, J., and McAuley, J. Justifying recommendations using distantly-labeled reviews and fine-grained aspects. In Proceedings of the 2019 conference on empirical methods in natural language processing and the 9th international joint conference on natural language process...

    2019

  36. [44]

    BLEU : a method for automatic evaluation of machine translation

    Papineni, K., Roukos, S., Ward, T., and Zhu, W.-J. BLEU : a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting of the Association for Computational Linguistics, pp.\ 311--318, 2002

    2002

  37. [45]

    Unbalanced diffusion Schr\"odinger bridge

    Pariset, M., Hsieh, Y.-P., Bunne, C., Krause, A., and De Bortoli, V. Unbalanced diffusion Schr\"odinger bridge. arXiv preprint arXiv:2306.09099, 2023

    2023 arXiv

  38. [46]

    and Xie, S

    Peebles, W. and Xie, S. Scalable diffusion models with transformers. In Proceedings of the IEEE/CVF international conference on computer vision, pp.\ 4195--4205, 2023

    2023

  39. [47]

    Diffusion bridge mixture transports, Schr \"o dinger bridge problems and generative modeling

    Peluchetti, S. Diffusion bridge mixture transports, Schr \"o dinger bridge problems and generative modeling. Journal of Machine Learning Research, 24 0 (374): 0 1--51, 2023

    2023

  40. [48]

    Computational optimal transport

    Peyr \'e , G., Cuturi, M., et al. Computational optimal transport. Foundations and Trends in Machine Learning , 11 0 (5-6): 0 355--607, 2019

    2019

  41. [49]

    and Niles-Weed, J

    Pooladian, A.-A. and Niles-Weed, J. Entropic estimation of optimal transport maps. arXiv preprint arXiv:2109.12004, 2021

    2021 arXiv

  42. [50]

    Prabhumoye, S., Tsvetkov, Y., Salakhutdinov, R., and Black, A. W. Style transfer through back-translation. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp.\ 866--876, 2018

    2018

  43. [51]

    DeFoG : Discrete flow matching for graph generation

    Qin, Y., Madeira, M., Thanou, D., and Frossard, P. DeFoG : Discrete flow matching for graph generation. arXiv preprint arXiv:2410.04263, 2024

    2024 arXiv

  44. [52]

    Language models are unsupervised multitask learners

    Radford, A., Wu, J., Child, R., Luan, D., Amodei, D., Sutskever, I., et al. Language models are unsupervised multitask learners. OpenAI blog, 1 0 (8): 0 9, 2019

    2019

  45. [53]

    High-resolution image synthesis with latent diffusion models

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., and Ommer, B. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 10684--10695, 2022

    2022

  46. [54]

    U-net: Convolutional networks for biomedical image segmentation

    Ronneberger, O., Fischer, P., and Brox, T. U-net: Convolutional networks for biomedical image segmentation. In Medical image computing and computer-assisted intervention--MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part III 18, ...

    2015

  47. [55]

    Improved techniques for training GAN s

    Salimans, T., Goodfellow, I., Zaremba, W., Cheung, V., Radford, A., and Chen, X. Improved techniques for training GAN s. In Advances in neural information processing systems, pp.\ 2234--2242, 2016

    2016

  48. [56]

    o dinger, E. \

    Schr \"o dinger, E. \"U ber die Umkehrung der Naturgesetze . Verlag der Akademie der Wissenschaften in Kommission bei Walter De Gruyter u. Company, 1931

    1931

  49. [57]

    Style transfer from non-parallel text by cross-alignment

    Shen, T., Lei, T., Barzilay, R., and Jaakkola, T. Style transfer from non-parallel text by cross-alignment. Advances in neural information processing systems, 30, 2017

    2017

  50. [58]

    D., Campbell, A., and Doucet, A

    Shi, Y., Bortoli, V. D., Campbell, A., and Doucet, A. Diffusion Schr\"odinger bridge matching. In Thirty-seventh Conference on Neural Information Processing Systems, 2023. URL https://openreview.net/forum?id=qy07OHsJT5

    2023

  51. [59]

    Y., Malkin, N., Fatras, K., Atanackovic, L., Zhang, Y., Huguet, G., Wolf, G., and Bengio, Y

    Tong, A. Y., Malkin, N., Fatras, K., Atanackovic, L., Zhang, Y., Huguet, G., Wolf, G., and Bengio, Y. Simulation-free Schr \"o dinger bridges via score and flow matching. In International Conference on Artificial Intelligence and Statistics, pp.\ 1279--1287. PMLR, 2024

    2024

  52. [60]

    Neural discrete representation learning

    Van Den Oord, A., Vinyals, O., et al. Neural discrete representation learning. Advances in neural information processing systems, 30, 2017

    2017

  53. [61]

    Solving Schr \"o dinger bridges via maximum likelihood

    Vargas, F., Thodoroff, P., Lamacraft, A., and Lawrence, N. Solving Schr \"o dinger bridges via maximum likelihood. Entropy, 23 0 (9): 0 1134, 2021

    2021

  54. [62]

    Digress: Discrete denoising diffusion for graph generation

    Vignac, C., Krawczuk, I., Siraudin, A., Wang, B., Cevher, V., and Frossard, P. Digress: Discrete denoising diffusion for graph generation. arXiv preprint arXiv:2209.14734, 2022

    2022 arXiv

  55. [63]

    Controllable unsupervised text attribute transfer via editing entangled latent representation

    Wang, K., Hua, H., and Wan, X. Controllable unsupervised text attribute transfer via editing entangled latent representation. Advances in Neural Information Processing Systems, 32, 2019

    2019

  56. [64]

    Energy-based diffusion language models for text generation

    Xu, M., Geffner, T., Kreis, K., Nie, W., Xu, Y., Leskovec, J., Ermon, S., and Vahdat, A. Energy-based diffusion language models for text generation. arXiv preprint arXiv:2410.21357, 2024

    2024 arXiv

  57. [65]

    A., Shechtman, E., and Wang, O

    Zhang, R., Isola, P., Efros, A. A., Shechtman, E., and Wang, O. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 586--595, 2018

    2018

  58. [66]

    Zhu, J.-Y., Park, T., Isola, P., and Efros, A. A. Unpaired image-to-image translation using cycle-consistent adversarial networks. In Proceedings of the IEEE international conference on computer vision, pp.\ 2223--2232, 2017

    2017

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.