Pith. sign in

REVIEW 3 major objections 4 minor 39 references

Latent Schrodinger Bridge: Prompting Latent Diffusion for Fast Unpaired Image-to-Image Translation

T0 review · 3 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper derives a Schrodinger-bridge ODE whose velocity splits into source, target, and noise predictors, and shows that one pretrained Stable Diffusion model can supply all three for fast unpaired image-to-image translation.

desk verdict Fast unpaired I2I via pretrained SD is real; the SB theory is an overclaim, not a proof. read the letter →

arxiv 2411.14863 v1 pith:YLUFRCQY submitted 2024-11-22 cs.CV cs.AIcs.LG

Jeongsol Kim , Beomsu Kim , Jong Chul Ye This is my paper · ORCID
classification cs.CVcs.AIcs.LG
keywords Schrodingerbridgeunpairedimage-to-imagetranslationprobabilityflowODElatentdiffusionStabletextualinversionSNRmatchingfastsampling
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 tries to establish that unpaired image-to-image translation can be made fast by reusing a pretrained latent diffusion model as an approximate Schrodinger bridge sampler. It derives a decomposition of the Schrodinger-bridge probability-flow ODE velocity into three interpretable predictors, and then argues that all three can be obtained from a single Stable Diffusion model via prompt optimization and an SNR-matching change of variables. If the construction is right, an image can be translated between domains in as few as eight neural function evaluations, with quality competitive to diffusion baselines that need many more steps.

What carries the argument

The central object is the decomposed Schrodinger-bridge ODE velocity v(xt,t) = ((1/2-t) $\sqrt$(tau) / $\sqrt$(t(1-t))) epsilon_hat(xt) + xhat_1(xt) - xhat_0(xt), where each term is a conditional expectation: the source predictor xhat_0, the target predictor xhat_1, and the noise predictor epsilon_hat. Proposition 1 provides the load-bearing justification that any coupling between P0 and P1 yields a valid probability-flow ODE between the marginals. The SNR-matching map ys = xt / $\sqrt$($sigma_t^{2}$ + 1) aligns the bridge state with what the VP diffusion denoiser saw during training, and prompt optimization via textual inversion turns one Stable Diffusion model into both the source and target predictors.

What would settle it

Construct a synthetic pair of distributions P0 and P1 where the posterior means E[x0|xt], E[x1|xt], and E[epsilon|xt] can be computed in closed form, then integrate the LSB ODE with the exact predictors and compare the terminal distribution with P1; a mismatch would invalidate the general-predictor claim. Separately, to test the SNR-matching premise, draw xt from the bridge construction with known x0, x1, and epsilon, and compare the VP denoiser output with the empirical conditional average over many samples: large systematic disagreement means the approximation fails.

Watch

Extended reading notes

Core claim

The central claim is that the Schrodinger-bridge probability-flow ODE velocity has the form v(xt,t) = ((1/2-t) $\sqrt$(tau) / $\sqrt$(t(1-t))) epsilon_hat(xt) + xhat_1(xt) - xhat_0(xt), and that this ODE still transports samples between P0 and P1 even if the three predictors are not the true Schrodinger-bridge posteriors. The source predictor xhat_0 repels from the source domain, the target predictor xhat_1 attracts toward the target domain, and the noise predictor epsilon_hat adds or removes noise depending on whether t < 0.5 or t > 0.5. The paper asserts that a variance-preserving diffusion denoiser trained on source images, after rescaling the bridge state by ys = xt / $\sqrt$($sigma_t^{2}$ + 1), provides valid estimates for x0 and epsilon, and the same trick with target-domain prompts provides x1. Prompt-optimized text embeddings let one Stable Diffusion model act as both domain predictors, and the result is claimed to be competitive unpaired translation with as few as 8 NFEs.

Load-bearing premise

The method assumes that a denoiser trained on source images can, after the SNR-matching rescaling, return an accurate estimate of the original source image from the bridge state, and that the same holds for target images; the paper does not prove this.

Editorial extensions

If this is right

  • Unpaired image-to-image translation can run with as few as 8 neural function evaluations using a single pretrained diffusion model, avoiding the separate inversion stage that slows down dual-diffusion baselines.
  • The source-repelling term addresses a failure mode of SDEdit, where incomplete translation leaves source structure intact because there is no force pushing away from the source domain.
  • Because any coupling between the source and target distributions suffices, expensive Schrodinger-bridge training is not needed to obtain a valid transport ODE.
  • Textual inversion with fewer than a thousand domain images is enough to create usable source and target predictors in Stable Diffusion.
  • The same ODE can translate in reverse by integrating from t = 1 to t = 0, giving bidirectional translation from one model and one set of optimized prompts.

Reading between the lines

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

  • The three-predictor decomposition is not tied to Stable Diffusion specifically: any score-based or flow-based denoiser could in principle supply the predictors, so the architecture may extend to faster distilled models or other latent diffusion backbones.
  • The SNR-matching argument aligns only the marginal noise level, not the full posterior distribution, so the reliability of the method on new domains should be tested by comparing denoiser outputs with exact posterior means on controlled data.
  • The dependence on textual inversion implies that domains far outside the pretrained model's prior may require finetuning or more than 1k images, a regime the paper does not explore.
  • Combining the LSB ODE with fast ODE solvers or learned step-size schedules might push the NFE count below 8 without the quality loss the paper reports at 4 NFEs.
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 proposes Latent Schrödinger Bridge (LSB), a method for fast unpaired image-to-image translation that leverages a pre-trained Stable Diffusion model. The authors decompose the SB probability flow ODE velocity into three terms—source predictor, target predictor, and noise predictor—and approximate each using a single text-conditioned VP diffusion denoiser, with domain-specific text embeddings obtained via textual inversion. They introduce an SNR-matching change of variables, a time-dependent noise predictor, CFG scaling, and a final denoising step. Experiments on Cat→Dog, Horse→Zebra, and Dog→Wild show improved FID/DINOv2 FD at 8 NFEs compared to DDIB and SDEdit, with additional ablations and examples across several domains.

Significance. If the theoretical grounding held, the paper would be a valuable practical contribution: it demonstrates that a single pre-trained diffusion model, without training an SB, can produce competitive unpaired translation in very few NFEs, and the component analysis (source/target/noise predictors) is intuitive. The empirical study is extensive and the code release is promised. However, the central theoretical claim—that VP diffusion denoisers compute the required SB posterior means after SNR matching—is not established and is in fact questionable. The paper itself concedes in Appendix G that no global OT map is constructed, so the proposed ODE is not a Schrödinger bridge in the strict sense. Thus the significance as a principled SB-based method is weakened; it may still be a useful empirical heuristic, but the current framing overstates the theoretical contribution.

major comments (3)
  1. [Section 4.2, Eq. (20)] The SNR-matching change of variables does not imply that a VP denoiser trained on P0 computes the SB conditional expectation E[x0|xt] required by Eq. (14). Under the LSB forward model, the rescaled input y_s = sqrt(alpha_s) x_t = sqrt(alpha_s)[(1-t)x0 + t x1] + sqrt(1-alpha_s) epsilon contains a structured signal from the target domain (the term t sqrt(alpha_s) x1), which is not present in the VP training distribution y_s = sqrt(alpha_s) x0 + sqrt(1-alpha_s) epsilon with x0 ~ P0. Matching only the scalar SNR does not align the conditional posteriors. For instance, with zero-mean isotropic Gaussian data and independent coupling, the true posterior E[x0|xt] differs from the Tweedie estimate produced by a VP denoiser for all t in (0,1). This undermines the premise of Proposition 1, which requires the predictors to be posterior means under a coupling with marginals P0 and P1. The Appendix G admission that the global OT map P^tau_01 is not constructed means the ODE's final marginal is not guaranteed to be P1. The authors should either provide a rigorous justification for why the VP denoiser approximation is valid (e.g., under additional assumptions or with a correction term), or reframe the method as an empirical approximation and validate the approximation directly (e.g., by measuring the discrepancy between the predictors and the true posterior means on synthetic data).
  2. [Algorithm 1, line 8] The pseudo-code writes y_{s_i} <- x_{t_i} / sqrt(alpha_{s_i}), which is the inverse of the formula y_s = sqrt(alpha_s) x_t given in Eq. (20). Since alpha_{s_i} = 1/(sigma_{t_i}^2 + 1), the line as written scales the input by sqrt(sigma_{t_i}^2+1), increasing the noise component rather than decreasing it. This is not a harmless typo: it directly contradicts the derivation in Section 4.2 and, if actually executed, would likely degrade performance substantially. The authors must correct this line (it should be y_{s_i} <- sqrt(alpha_{s_i}) x_{t_i}) and clarify whether the reported experiments used the corrected formula.
  3. [Section 5.1 and Appendix C.5] The hyperparameters t0, tau, and the CFG scale omega are selected by grid search on the same tasks and metrics (FID) that are reported in the main results (Table 1 and Figure 3). This makes the reported improvements partly a product of tuning rather than a general property of the algorithm. To support the claim of broad applicability, the authors should either report results with a fixed, task-independent hyperparameter setting, or show sensitivity curves and report the performance on held-out tasks or with a validation split that is not used for tuning. Otherwise, the reader cannot assess how much of the advantage is due to the method versus the grid search.
minor comments (4)
  1. [Appendix G] The limitations paragraph states that the method is not a true Schrödinger bridge because the global OT map is not constructed. This admission is in tension with the title and abstract, which present LSB as a Schrödinger bridge method. The authors should either soften the claim or rename the method to avoid overclaiming.
  2. [Section 4.1, Eq. (12)] The derivation of Eq. (12) from Eq. (11) uses sigma_t = sqrt(t(1-t) tau), but the expression has a typo: the factor in the noise term should read (1/2-t) sqrt(tau) / sqrt(t(1-t)), not (1/2-t) sqrt(tau) / sqrt(t(1-t)) with the square root misplaced. Please correct the formula and re-check the subsequent equations.
  3. [Section 4.3, 'Prompt optimization'] The text says 'we can improve the reflectance of text prompt'—likely intended as 'reflectance' or 'relevance'? Please clarify the wording.
  4. [References] Some references appear incomplete or have incorrect page numbers (e.g., [12] is cited as 'arXiv preprint' without a year, and [30] has an extra '2' at the end). Please ensure all references are accurately formatted.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the LSB ODE is a flow-matching construction with an explicitly approximate VP-denoiser substitution, validated against external baselines.

full rationale

The paper's derivation chain is not circular. The SB ODE velocity decomposition (Eqs. 7-12) is a direct application of linearity of expectation and the reparametrization trick, and Proposition 1 is proven by invoking Theorem 3.1 of an external flow-matching reference ([32]); the marginals P0/P1 at t=0/1 follow from the construction of Pt in Eq. (13) using a coupling Γ01 with those marginals. This is a standard flow-matching construction, not a conclusion assumed in its own premises. The later use of a pretrained Stable Diffusion VP denoiser (Sec. 4.2) is explicitly an approximation: the paper states that xt and ys have different formulations, matches only their SNRs, and does not claim the VP posterior equals the SB posterior. Appendix G honestly concedes that the global OT map Pτ_01 is not constructed, so the method is not literally a Schrödinger bridge and no uniqueness/self-citation chain is used to force the choice. The only self-citations ([12], [13]) appear in a sanity-check proof and related-work discussion, respectively, and are not load-bearing for the central I2I claim. The hyperparameters τ, t0, and ω are disclosed as grid-searched in Appendix C.5 on the same tasks; this is standard tuning, not a fitted parameter renamed as a prediction. The headline empirical claims are benchmarked against external baselines (DDIB, PnP, SDEdit) with reported FID/KD/DINOv2 FD scores, so the empirical content is self-contained. One non-circular caveat is that Algorithm 1 line 8 writes y_s ← x_t / sqrt(α_s), which inverts Eq. (20); this is an implementation typo or inconsistency rather than a circular reduction, and a correctness risk only.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on three items not paid for upstream: the standard flow-matching theorem that the paper cites, and two ad hoc assumptions about how Stable Diffusion's VP scores approximate SB posterior means. The free parameters are hyperparameters selected by grid search and ablation on the exact tasks and metrics used for the headline results.

free parameters (4)
  • tau (SB variance) = sqrt(tau) = 2.5 for main 8-NFE runs; 1.5, 2.3, 3.0, 3.5 for other NFE settings
    Controls the noise level in the bridge trajectory. Selected by ablation and grid search in Appendix C.5; directly affects translation quality and blur.
  • initial SB time t0 = 0.2
    Chosen via ablation (Figure 7). Larger t0 blurs output, smaller t0 retains too much source content.
  • CFG scale omega = 11.0 in main experiments; grid-searched values 3.0, 5.0, 11.0 for different settings
    Classifier-free guidance weight that controls how strongly the target text prompt steers xhat_1. Selected per task and NFE in Appendix C.5.
  • final denoising step = always applied
    A single VP denoising step after the ODE solution. Ablation in Table 2 shows it improves FID; used in all main experiments.
assumptions (5)
  • standard math Thm 3.1 of Tong et al. (2023): for any coupling Gamma_01 and any conditional path P_{t|01}, the regression objective yields the marginal probability-flow ODE.
    Used without proof as the basis of Proposition 1; cited in Appendix B.1 and Section 3.
  • ad hoc to paper Text-conditioned Stable Diffusion VP denoisers, after textual inversion on roughly 1k images, compute the posterior means in Eq. (14) for the source and target domains.
    The central approximation of the paper. No proof that the SB posterior E[x0 | xt] equals the VP Tweedie estimate on y_s; invoked in Section 4.3 and Algorithm 1.
  • ad hoc to paper Matching the SNR between xt and ys is sufficient to make a VP score model a valid predictor for SB variables, even though the clean signal in xt is mu_t = (1-t)x0 + t x1 rather than a draw from P0.
    Section 4.2, Eq. (20). The distribution mismatch of the clean signal is not analyzed.
  • domain assumption Stable Diffusion's latent space has enough capacity and semantic prior for all tested domain shifts.
    The method operates in SD's latent space; the paper does not analyze failure modes for domains poorly covered by the pretrained prior.
  • domain assumption Euler integration of the LSB ODE with N=8 steps, plus a final denoising step, accurately approximates the intended SB-like probability path.
    The paper does not bound the discretization error; the choice of N and the final denoising step are justified mostly by experiments in Section 5 and Appendix C.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Latent Schrodinger Bridge: Prompting Latent Diffusion for Fast Unpaired Image-to-Image Translation." pith.science (2026). https://pith.science/paper/YLUFRCQY

@misc{pith2026241114863,
  author       = {Pith},
  title        = {Pith review of: Latent Schrodinger Bridge: Prompting Latent Diffusion for Fast Unpaired Image-to-Image Translation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YLUFRCQY}},
  note         = {Machine review of arXiv:2411.14863}
}
read the original abstract

Diffusion models (DMs), which enable both image generation from noise and inversion from data, have inspired powerful unpaired image-to-image (I2I) translation algorithms. However, they often require a larger number of neural function evaluations (NFEs), limiting their practical applicability. In this paper, we tackle this problem with Schrodinger Bridges (SBs), which are stochastic differential equations (SDEs) between distributions with minimal transport cost. We analyze the probability flow ordinary differential equation (ODE) formulation of SBs, and observe that we can decompose its vector field into a linear combination of source predictor, target predictor, and noise predictor. Inspired by this observation, we propose Latent Schrodinger Bridges (LSBs) that approximate the SB ODE via pre-trained Stable Diffusion, and develop appropriate prompt optimization and change of variables formula to match the training and inference between distributions. We demonstrate that our algorithm successfully conduct competitive I2I translation in unsupervised setting with only a fraction of computation cost required by previous DM-based I2I methods.

Figures

Figures reproduced from arXiv: 2411.14863 by the authors.

Figure 1
Figure 1. Overview of the proposed method. (A) Using an unpaired dataset, we optimize text embeddings to obtain conditioned score functions using Stable Diffusion 1.5. (B) By decomposing Schrodinger Bridge (LSB) ODE with these optimized score functions, (C) we ¨ achieve high-quality image-to-image translation with as few as 8 NFEs. Abstract Diffusion models (DMs), which enable both image genera￾tion from noise and inversion f… view at source ↗
Figure 2
Figure 2. Decomposition of different ODEs for image to image translation. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. FID vs. NFE for three image translation tasks. [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison for Cat2Dog (8 NFEs) [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 7
Figure 7. Figure 7: As t0 increases to 1, the resulting images become progressively blurrier and lose details from the source im￾age (see last row). This is because we initialize xt0 accord￾ing to Eq. (21), and a larger t0 results in a significant SNR mismatch between xt and the training …
Figure 8
Figure 8. Figure 8: Uncurated samples generated by VP diffusion with opti [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Quantitative comparison across NFEs: We evaluate FID, using DINOv2 as the encoder, and LPIPS on translated images from [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]
Figure 10
Figure 10. Figure 10: Translated image without and with SNR matching. [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 12
Figure 12. Figure 12: Image translation example by LSB for photo [PITH_FULL_IMAGE:figures/full_fig_p015_12.png]
Figure 13
Figure 13. Figure 13: Image translation example by LSB for male [PITH_FULL_IMAGE:figures/full_fig_p016_13.png]
Figure 14
Figure 14. Figure 14: Image translation example by LSB for bibimbap [PITH_FULL_IMAGE:figures/full_fig_p017_14.png]
Figure 15
Figure 15. Figure 15: Image translation example by LSB for pulled-pork sandwich [PITH_FULL_IMAGE:figures/full_fig_p018_15.png]
Figure 16
Figure 16. Figure 16: Image translation example by LSB for apple [PITH_FULL_IMAGE:figures/full_fig_p019_16.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 34 canonical work pages

  1. [1]

    Demystifying mmd gans

    Mikołaj Bi ´nkowski, Danica J Sutherland, Michael Arbel, and Arthur Gretton. Demystifying mmd gans. arXiv preprint arXiv:1801.01401, 2018. 7, 2

    arXiv 2018

  2. [2]

    Diffusion Schr ¨odinger Bridge with Appli- cations to Score-Based Generative Modeling

    Valentin De Bortoli, James Thornton, Jeremy Heng, and Ar- naud Doucet. Diffusion Schr ¨odinger Bridge with Appli- cations to Score-Based Generative Modeling. In NeurIPS,

    work page

  3. [3]

    Mccann, Marc L

    Hyungjin Chung, Jeongsol Kim, Michael T. Mccann, Marc L. Klasky, and Jong Chul Ye. Diffusion Posterior Sam- pling for General Noisy Inverse Problems. In ICLR, 2023. 1

    work page 2023

  4. [4]

    Diffusion Models Beat GANs on Image Synthesis

    Prafulla Dhariwal and Alex Nichol. Diffusion Models Beat GANs on Image Synthesis. In NeurIPS, 2021. 1

    work page 2021

  5. [5]

    Bermano, Gal Chechik, and Daniel Cohen-Or

    Rinon Gal, Yuval Alaluf, Yuval Atzmon, Or Patashnik, Amit H. Bermano, Gal Chechik, and Daniel Cohen-Or. An image is worth one word: Personalizing text-to-image gen- eration using textual inversion, 2022. 6, 2

    work page 2022

  6. [6]

    Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio

    Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative Adversarial Nets. In NeurIPS,

    work page

  7. [7]

    Adversarial Schr¨odinger Bridge Matching

    Nikita Gushchin, Daniil Selikhanovych, Sergei Kholkin, Evgeny Burnaev, and Alexander Korotin. Adversarial Schr¨odinger Bridge Matching. In NeurIPS, 2024. 3

    work page 2024

  8. [8]

    GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, and Bernhard Nessler. GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium. In NeurIPS, 2017. 7, 2

    work page 2017

Show all 39 references
  1. [9]

    Denoising Dif- fusion Probabilistic Models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising Dif- fusion Probabilistic Models. In NeurIPS, 2020. 1

    2020

  2. [10]

    Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A. Efros. Image-to-Image Translation with Conditional Adver- sarial Networks. In CVPR, 2017. 2

    2017

  3. [11]

    Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Ge- ometry and Semantics

    Alex Kendall, Yarin Gal, and Roberto Cipolla. Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Ge- ometry and Semantics. In CVPR, 2018. 7, 2

    2018

  4. [12]

    Sim- ple reflow: Improved techniques for fast flow models

    Beomsu Kim, Yu-Guan Hsieh, Michal Klein, Marco Cuturi, Jong Chul Ye, Bahjat Kawar, and James Thornton. Sim- ple reflow: Improved techniques for fast flow models. arXiv preprint arXiv:2410.07815, 2024. 1

    2024 arXiv

  5. [13]

    Unpaired Image-to-Image Translation via Neural Schr¨odinger Bridge

    Beomsu Kim, Gihyun Kwon, Kwanyoung Kim, and Jong Chul Ye. Unpaired Image-to-Image Translation via Neural Schr¨odinger Bridge. In ICLR, 2024. 3, 4

    2024

  6. [14]

    Noise2Score: Tweedie’s Approach to Self-Supervised Image Denoising without Clean Images

    Kwanyoung Kim and Jong Chul Ye. Noise2Score: Tweedie’s Approach to Self-Supervised Image Denoising without Clean Images. In NeurIPS, 2021. 5

    2021

  7. [15]

    Kingma and Max Welling

    Diederik P. Kingma and Max Welling. Auto-Encoding Vari- ational Bayes. In ICLR, 2014. 4

    2014

  8. [16]

    Yaron Lipman, Ricky T. Q. Chen, Heli Ben-Hamu, Maximil- ian Nickel, and Matthew Le. Flow Matching for Generative Modeling. In ICLR, 2023. 3

    2023

  9. [17]

    Theodorou, and Ricky T

    Guan-Horng Liu, Yaron Lipman, Maximilian Nickel, Brian Karrer, Evangelos A. Theodorou, and Ricky T. Q. Chen. Generalized Schr ¨odinger Bridge Matching. In ICLR, 2024. 3

    2024

  10. [18]

    Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow

    Xingchao Liu, Chengyue Gong, and Qiang Liu. Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow. arXiv preprint arXiv:2209.03003 ,

    arXiv

  11. [19]

    DPM-Solver: A Fast ODE Solver for Dif- fusion Probabilistic Model Sampling in Around 10 Steps

    Cheng Lu, Yuhao Zhou, Fan Bao, Jianfei Chen, Chongxuan Li, and Jun Zhu. DPM-Solver: A Fast ODE Solver for Dif- fusion Probabilistic Model Sampling in Around 10 Steps. In NeurIPS, 2022. 2, 3

    2022

  12. [20]

    SDEdit: Guided Im- age Synthesis and Editing with Stochastic Differential Equa- tions

    Chenlin Meng, Yutong He, Yang Song, Jiaming Song, Jiajun Wu, Jun-Yan Zhu, and Stefano Ermon. SDEdit: Guided Im- age Synthesis and Editing with Stochastic Differential Equa- tions. In ICLR, 2022. 2, 6

    2022

  13. [21]

    Efros, Richard Zhang, and Jun- Yan Zhu

    Taesung Park, Alexei A. Efros, Richard Zhang, and Jun- Yan Zhu. Contrastive Learning for Unpaired Image-to-Image Translation. In ECCV, 2020. 2

    2020

  14. [22]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

    2021

  15. [23]

    High-Resolution Image Synthesis with Latent Diffusion Models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-Resolution Image Synthesis with Latent Diffusion Models. In CVPR, 2022. 1

    2022

  16. [24]

    Progressive Distillation for Fast Sampling of Diffusion Models

    Tim Salimans and Jonathan Ho. Progressive Distillation for Fast Sampling of Diffusion Models. In ICLR, 2022. 3

    2022

  17. [25]

    Diffusion Schr ¨odinger Bridge Matching

    Yuyang Shi, Valentin De Bortoli, Andrew Campbell, and Ar- naud Doucet. Diffusion Schr ¨odinger Bridge Matching. In NeurIPS, 2023. 2, 3

    2023

  18. [26]

    Denois- ing Diffusion Implicit Models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denois- ing Diffusion Implicit Models. In ICLR, 2021. 1, 2, 3

    2021

  19. [27]

    Improved Techniques for Training Consistency Models

    Yang Song and Prafulla Dhariwal. Improved Techniques for Training Consistency Models. In ICLR, 2024. 2, 3

    2024

  20. [28]

    Kingma, Ab- hishek Kumar, Stefano Ermon, and Ben Poole

    Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Ab- hishek Kumar, Stefano Ermon, and Ben Poole. Score-Based Generative Modeling through Stochastic Differential Equa- tions. In ICLR, 2021. 1, 2, 3

    2021

  21. [29]

    Consistency Models

    Yang Song, Prafulla Dhariwal, Mark Chen, and Ilya Sutskever. Consistency Models. In ICML, 2023. 2, 3

    2023

  22. [30]

    Exposing flaws of generative model evaluation metrics and their un- fair treatment of diffusion models

    George Stein, Jesse Cresswell, Rasa Hosseinzadeh, Yi Sui, Brendan Ross, Valentin Villecroze, Zhaoyan Liu, Anthony L Caterini, Eric Taylor, and Gabriel Loaiza-Ganem. Exposing flaws of generative model evaluation metrics and their un- fair treatment of diffusion models. In Advan...

    2023

  23. [31]

    Dual Diffusion Implicit Bridges for Image-to-Image Trans- lation

    Xuan Su, Jiaming Song, Chenlin Meng, and Stefano Ermon. Dual Diffusion Implicit Bridges for Image-to-Image Trans- lation. In ICLR, 2023. 2, 6

    2023

  24. [32]

    Improving and generalizing flow-based generative models with minibatch optimal transport

    Alexander Tong, Kilian Fatras, Nikolay Malkin, Guillaume Huguet, Yanlei Zhang, Jarrid Rector-Brooks, Guy Wolf, and Yoshua Bengio. Improving and generalizing flow-based generative models with minibatch optimal transport. arXiv preprint arXiv:2302.00482, 2023. 3, 1

    2023 arXiv

  25. [33]

    Splicing vit features for semantic appearance transfer

    Narek Tumanyan, Omer Bar-Tal, Shai Bagon, and Tali Dekel. Splicing vit features for semantic appearance transfer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10748–10757, 2022. 7, 2 9

    2022

  26. [34]

    Plug-and-Play Diffusion Features for Text-Driven Image-to-Image Translation

    Narek Tumanyan, Michal Geyer, Shai Bagon, and Tali Dekel. Plug-and-Play Diffusion Features for Text-Driven Image-to-Image Translation. In CVPR, 2023. 2, 6

    2023

  27. [35]

    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 ICML, 2021. 2

    2021

  28. [36]

    Imagere- ward: Learning and evaluating human preferences for text- to-image generation

    Jiazheng Xu, Xiao Liu, Yuchen Wu, Yuxuan Tong, Qinkai Li, Ming Ding, Jie Tang, and Yuxiao Dong. Imagere- ward: Learning and evaluating human preferences for text- to-image generation. Advances in Neural Information Pro- cessing Systems, 36, 2024. 7, 2

    2024

  29. [37]

    Fast Sampling of Dif- fusion Models with Exponential Integrator

    Qinsheng Zhang and Yongxin Chen. Fast Sampling of Dif- fusion Models with Exponential Integrator. In ICLR, 2023. 2, 3

    2023

  30. [38]

    Fast ODE-based Sampling for Diffusion Models in Around 5 Steps

    Zhenyu Zhou, Defang Chen, Can Wang, and Chun Chen. Fast ODE-based Sampling for Diffusion Models in Around 5 Steps. In CVPR, 2024. 2, 3

    2024

  31. [39]

    !” to represent the source domain and “*

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A. Efros. Unpaired Image-to-Image Translation using Cycle- Consistent Adversarial Networks. In ICCV, 2017. 2 10 Latent Schr¨odinger Bridge: Prompting Latent Diffusion for Fast Unpaired Image-to-Image Translation Supplementar...

    2017

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.