Pith. sign in

REVIEW 3 major objections 4 minor 3 cited by

On the Guidance of Flow Matching

T0 review · 3 major / 4 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read Flow matching guidance reduces to a single explicit vector-field identity that covers exact, learned, and classical diffusion methods.

desk verdict A genuinely unifying framework for flow-matching guidance, with a load-bearing approximation that needs to be stated more carefully before the exactness claims are accepted. read the letter →

arxiv 2502.02150 v3 pith:NNRWHVJ5 submitted 2025-02-04 cs.LG

Ruiqi Feng , Chenglei Yu , Wenhao Deng , Peiyan Hu , Tailin Wu This is my paper · ORCID
classification cs.LG
keywords flowmatchingenergyguidancevectorfieldMonteCarloestimationtraining-baseddiffusionofflinereinforcementlearninggenerativemodeling
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

This paper claims that steering a flow-matching generative model—one that pushes a source distribution to a target by following a learned vector field—toward the energy-weighted distribution $p'(x) \propto p(x) e^{-J(x)}$ reduces to adding one explicit guidance vector field to the learned field. The identity, Theorem 3.1, applies to arbitrary source distributions, couplings, and conditional paths, not only to the Gaussian-source setting where flow matching becomes diffusion. If the identity is correct, many classical diffusion guidance methods are special cases of one formula, and new training-free and training-based guidance methods follow from it. The paper tests the framework on synthetic distributions, image inverse problems, and offline reinforcement learning planning.

What carries the argument

The carrying object is the guidance vector field identity itself: a closed-form expression for the additive field that reweights the terminal distribution by the Boltzmann factor. The mechanism that keeps the identity tractable is the reverse coupling ratio $P = \pi'(x_0|x_1)/\pi(x_0|x_1)$, which is exactly 1 for independent couplings; every practical method in the paper operates under the $P=1$ approximation. Under the affine Gaussian uncoupled path, conditional vector fields become conditional scores, which is what turns the identity into the gradient guidance formula used in diffusion models.

What would settle it

On a low-dimensional dataset with strongly dependent mini-batch OT couplings and a sharply varying energy $J$, guide samples with $g^{MC}$ using many Monte Carlo samples and compare the terminal distribution to $p(x)e^{-J(x)}/Z$; a mismatch beyond sampling error shows that the $P=1$ premise fails for dependent couplings.

Watch

Extended reading notes

Core claim

At the center is the exact guidance identity of Theorem 3.1: adding $g_t(x_t)$ to the original vector field $v_t(x_t)$ generates the path ending at $p'(x)$, where $g_t(x_t) = \int \big(P e^{-J(x_1)}/Z_t(x_t) - 1\big) v_{t|z}(x_t|z) p(z|x_t) dz$ and $Z_t(x_t) = \int P e^{-J(x_1)} p(z|x_t) dz$, with $P$ the reverse coupling ratio. Under the uncoupled affine Gaussian path assumption the identity collapses onto the familiar diffusion guidance $\nabla_{x_t}\log Z_t(x_t)$ up to a schedule factor. From this single identity the paper derives asymptotically exact Monte Carlo guidance ($g^{MC}$), learned guidance ($g^{\phi}$), localized Taylor guidance ($g^{local}$, $g^{cov}$), and Gaussian-approximation guidance ($g^{sim}$), each covering previously known diffusion methods as special cases.

Load-bearing premise

The load-bearing premise is that the reverse coupling ratio $P$, the ratio between the guided and unguided distributions of the source given a target point, is 1; exact for independent couplings, this is only an approximation for dependent couplings with rapidly varying energy.

Editorial extensions

If this is right

  • Any pre-trained flow matching model with a non-Gaussian source or dependent coupling can be guided without retraining by $g^{MC}$, which is asymptotically exact as the Monte Carlo sample count grows.
  • Learned guidance $g^{\phi}$ can be trained with simple conditional losses whose gradients match an intractable marginal loss, giving exact training-based guidance for general flow matching.
  • Classical diffusion guidance methods—DPS, LGD, $\Pi$GDM, and OT-ODE—are recovered as special cases, so their validity extends to flow matching paths beyond the uncoupled affine Gaussian setting.
  • Approximate gradient guidance is reliable when the energy is smooth and the flow time is close to 1, while $g^{MC}$ excels in low-dimensional or frequently conditioned tasks and fails on image inverse problems because of Monte Carlo variance.

Reading between the lines

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

  • The framework leaves open the possibility of choosing the reverse coupling deliberately rather than approximating it, which would make exact guidance available for optimal-transport and rectified flows.
  • Using a learned approximate guidance as the proposal in $g^{MC}$ importance sampling (the paper's $g^{MC\text{-}IS}$ sketch) is a natural way to make asymptotically exact guidance practical in high dimensions.
  • If exact coupling-aware guidance becomes practical, flows with straighter paths, such as mini-batch OT or rectified flows, would bring guided sampling closer to one-step accuracy than Gaussian-source diffusion guidance can.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 a unified framework for energy-guided sampling with flow matching. Starting from conditional flow matching, Theorem 3.1 expresses the guidance field as a difference of conditional vector fields averaged under the reweighted coupling, with a reverse-coupling ratio P. The paper then sets P≈1 and derives: a training-free Monte Carlo guidance method gMC (Algorithm 1); localized approximations glocal/gcov and gsim that recover DPS, LGD, ΠGDM, and OT-ODE as special cases; and training-based guidance losses gϕ. Experiments cover 2D synthetic flows with non-Gaussian sources, D4RL locomotion planning, and CelebA-HQ inverse problems. Full proofs and an error bound for the localized approximation are provided in the appendices, and code is released.

Significance. If the framework holds as stated, this is a substantial unification: it would justify guidance for non-Gaussian source distributions, dependent couplings, and general conditional paths, and it would place several established diffusion guidance methods as special cases. The paper's strengths are the clean marginalization identity in Theorem 3.1, the explicit recovery of classical methods in Appendices A.4 and A.13, an error bound for glocal in Appendix A.9, extensive appendices, and publicly available code. The central claims are tempered by two substantial caveats: the P=1 approximation changes the source distribution for dependent couplings, and the finite-sample unbiasedness of the self-normalized Monte Carlo estimator is not correct as stated. These caveats affect the interpretation of the OT-CFM experiments and the ``applicable to any source distribution'' claims, but the core identity remains a sound construction for the reweighted-path vector field.

major comments (3)
  1. [§3.1, Theorem 3.1; Appendix A.3] Theorem 3.1 is a correct identity for the path p'_t(x_t)=∫p_t(x_t|z)π'(x0|x1)Z^{-1}p(x1)e^{-J(x1)}dz, but the guidance problem is solved by integrating v'_t from the original source p0 only when p'_0(x0)=p0(x0). Setting P=1 corresponds to π'=π, which gives p'_0(x0)=∫π(x0|x1)Z^{-1}p(x1)e^{-J(x1)}dx1, and this equals p0 only when the coupling is independent. All implemented methods (gMC, gcov, gsim, gϕ) run with P=1, so for mini-batch OT couplings Theorem 3.1 does not by itself establish exactness for sampling started from p0. Appendix A.3 acknowledges this, but Table 3 validates P≈1 only indirectly by comparing the vector fields of OT-CFM and CFM, not by measuring P or the source mismatch, and Section 5 states that the guidance direction is no longer correct under strong coupling and strong J. The abstract and Section 1 should be scoped accordingly, or the paper should directly measure P or the induced source mismatch in the strong-coupling regime.
  2. [§3.2, Algorithm 1] Section 3.2 states that gMC is ``unbiased and applicable to any source distribution.'' Algorithm 1, however, uses a self-normalized estimator: both \itilde Z_t and \itilde p_t are Monte Carlo estimates formed from the same samples, and gMC is a ratio of these estimates. A self-normalized importance-sampling estimator is generally biased at finite N and only asymptotically unbiased under standard regularity conditions. This affects the ``asymptotically exact'' claim and the interpretation of the scaling experiment in Figure 4. Please replace ``unbiased'' with ``asymptotically unbiased'' or provide a genuinely unbiased estimator, for example by using independent sample sets for the numerator and denominator or by using a known normalizing constant pt(xt).
  3. [§3.5, Appendix A.7] Section 3.5 claims that the losses ℓVGM, ℓRGM, and ℓMRGM produce the same minimizer as ℓGM and that all provide unbiased gradient estimates, but the proof in Appendix A.7 relies on the assumption in Eq. (51) that vθ(xt,t)=E_{z∼p(z|xt)}[v_{t|z}(xt|z)], i.e., a perfectly learned flow-matching vector field. For a practical pre-trained model with approximation error, the gradients of ℓVGM and ℓRGM differ from that of ℓGM by terms involving vθ−vt. This assumption should be stated in the main text, and the equivalence claims should be qualified or supplemented with an error bound that depends on the flow-matching training error.
minor comments (4)
  1. [Appendix A.3, Table 3] The sentence introducing Table 3 contains a typo: ``the VF of the OT-CFM (batch size 128) in and uncoupled CFM trained'' should read ``... in and uncoupled CFM trained...'' or similar, and Table 3 is not cited at the point in Section 3.1 where P=1 is first introduced.
  2. [§3.2, Eq. (3)] The notation E_{x1,x0∼p(z)} is confusing because p(z) is a joint coupling π(x0,x1); please write E_{(x0,x1)∼π} or equivalent so that the coupling is explicit.
  3. [Algorithm 1] In Algorithm 1 the symbol \itilde p_t is used both for the estimated marginal density and for the normalized importance weights; a short comment or a distinct variable name would clarify the self-normalization step.
  4. [Figure 4 caption] The caption is confusing: ``the guided generation errors (crosses) do not converge to the dashed lines is that they measure the W2 distance of p(x1) and p(x1)e^{−J(x1)}/Z'' mixes quantities. Please state explicitly which two distributions are compared for the dashed lines and which two for the crosses.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Theorem 3.1 is a constructive identity and the P=1 approximation is an acknowledged limitation, not a self-referential reduction.

full rationale

The paper's central derivation (Theorem 3.1, Eq. (1), Appendix A.2) is a constructive identity. The guided vector field v'_t is defined by conditional flow matching for the reweighted coupling p'(z)=π'(x0|x1)p(x1)e^{-J(x1)}/Z, and gt is then solved from gt=v'_t-vt. This is a rearrangement of the definition of v'_t; the nontrivial content is the explicit expression of gt in terms of the unguided posterior p(z|xt), not an empirical prediction fitted to data. The training-based losses in Section 3.5 are supervised regressions onto that defined gt via Proposition 3.5, and gMC in Section 3.2 is self-normalized importance sampling of Eq. (1); neither relabels a fitted parameter as a prediction. The P≈1 approximation is the main load-bearing assumption for dependent couplings, and the paper explicitly flags it: Section 5 states 'When the coupling is strong, the guidance VF no longer has the correct direction,' and Appendix A.3 discusses the induced source-distribution error. The empirical check in Appendix A.3/Table 3 compares OT-CFM and CFM vector fields rather than measuring P, which weakens validation but is an approximation gap, not circularity. The paper also benchmarks against external methods (DPS, LGD, PiGDM) and does not rely on a self-citation chain or an imported uniqueness theorem. Accordingly, no step reduces to its own inputs by construction.

Assumptions & free parameters 6 free parameters · 7 assumptions · 0 invented entities

The central derivation rests on the flow matching conditional vector field identity, plus several modeling assumptions. The most load-bearing is the P=1 approximation, which is exactly true only for independent couplings. The other assumptions (affine paths, localized or Gaussian conditional distributions, exact learned v_theta) are standard or explicitly approximate, but each limits the advertised generality. Six hyperparameters are fitted or tuned in experiments, which affects the empirical comparisons but not the core theoretical identity.

free parameters (6)
  • Guidance scale of J for gMC in planning = Tuned in {0.2, 1, 2, 3, 5}
    Appendix B.2 tunes the scale of the energy function J for gMC in D4RL experiments; this controls how strongly the distribution is reweighted.
  • Stability epsilon for gMC = Tuned values 1e-3, 5e-3, 1e-2, 5e-2 in ablation
    Table 6 varies epsilon and reports different scores; epsilon is a numerical stabilization hyperparameter.
  • gcov-A schedule lambda_t = Tuned among constant, cosine, exponential, linear decay with scaler in {0.01, 0.1, 1.0, 10.0}
    gcov-A uses a manually set covariance schedule; Appendix B.2 tunes its shape and scale per task.
  • gcov-G scale hyperparameter = Same tuning set as gcov-A
    Although the functional form of lambda_cov-G is fixed by the flow schedule, Appendix B.2 states the same schedule and scaler hyperparameters are tuned for gcov-G.
  • gsim-MC standard deviation and guidance scale = Std in {0.1, 0.5, 1, 10}; scale in {0.1, 1, 10}
    The assumed standard deviation of p(x1|xt) and the guidance scale are tuned in planning experiments (Appendix B.2).
  • gsim-inv-A hyperparameters lambda_t and r_t = Hyperparameters, exact values not enumerated in main text
    Eq. (9) defines lambda_t and r_t as hyperparameters for the inverse-problem guidance method.
assumptions (7)
  • ad hoc to paper Reverse coupling ratio P = pi'(x0|x1)/pi(x0|x1) is set to 1.
    Exactly true only for independent couplings; for dependent couplings it is an approximation validated by empirical vector field similarity in Table 3, and the paper admits it fails when coupling is strong and J varies intensively.
  • domain assumption Assumption 3.2: conditional probability path is affine, xt = alpha_t x1 + beta_t x0 + sigma_t epsilon, with sigma_t and sigma_dot_t small.
    Used to estimate x_hat_1 from the learned vector field and to derive gcov guidance. It restricts the generality of the derived approximations.
  • domain assumption Assumption 3.3: uncoupled affine Gaussian path, standard Gaussian source and independent coupling.
    Used for the Jacobian trick and to recover DPS, LGD, and PiGDM as special cases. This is exactly the setting where flow matching reduces to diffusion models.
  • domain assumption p(z|xt) is localized around its mean, justifying Taylor expansion of e^{-J}.
    Basis of glocal and gcov approximations; the error bound is only an upper bound and can be large for non-smooth J or high-variance p(x1|xt).
  • domain assumption p(z|xt) is approximated as a Gaussian in Section 3.4.
    Needed for gsim-MC and gsim-inv; the covariance is approximated or set as a hyperparameter, introducing bias.
  • ad hoc to paper v_theta is a perfectly learned vector field (Eq. 51) in the proof that VGM and RGM losses have the same gradient as the GM loss.
    The main text states these losses produce the same minimizer without flagging that the proof requires an exact v_theta.
  • domain assumption For gsim-inv-A, partial x_hat_1 / partial x_t is approximated as I.
    Appendix A.13 states this approximation is accurate only as t approaches 1, so it may be poor at early flow times.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On the Guidance of Flow Matching." pith.science (2026). https://pith.science/paper/NNRWHVJ5

@misc{pith2026250202150,
  author       = {Pith},
  title        = {Pith review of: On the Guidance of Flow Matching},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NNRWHVJ5}},
  note         = {Machine review of arXiv:2502.02150}
}
read the original abstract

Flow matching has shown state-of-the-art performance in various generative tasks, ranging from image generation to decision-making, where generation under energy guidance (abbreviated as guidance in the following) is pivotal. However, the guidance of flow matching is more general than and thus substantially different from that of its predecessor, diffusion models. Therefore, the challenge in guidance for general flow matching remains largely underexplored. In this paper, we propose the first framework of general guidance for flow matching. From this framework, we derive a family of guidance techniques that can be applied to general flow matching. These include a new training-free asymptotically exact guidance, novel training losses for training-based guidance, and two classes of approximate guidance that cover classical gradient guidance methods as special cases. We theoretically investigate these different methods to give a practical guideline for choosing suitable methods in different scenarios. Experiments on synthetic datasets, image inverse problems, and offline reinforcement learning demonstrate the effectiveness of our proposed guidance methods and verify the correctness of our flow matching guidance framework. Code to reproduce the experiments can be found at https://github.com/AI4Science-WestlakeU/flow_guidance.

Figures

Figures reproduced from arXiv: 2502.02150 by the authors.

Figure 1
Figure 1. Overview of guidance methods in the paper. We start with a unified guidance expression and derive different [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Results of the synthetic dataset with different source (blue) and target (red) distributions. We visualize the start/end [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. R distribution of generated trajectories in Loco￾motion. g MC matches the target gray dashed line well. The results demonstrate that ΠGDM is generally better on all tasks, being the best in 8 out of 12 metrics. g cov-G has a similar but slightly worse performance than ΠGDM, be￾ing the best or the runner-up in all 4 metrics of the super￾resolution task if ranking the results of CFM and OT-CFM separately and 3 out of … view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Error scaling with Monte Carlo sample number. In the synthetic dataset, the guidance performance ( [PITH_FULL_IMAGE:figures/full_fig_p033_4.png]
Figure 5
Figure 5. Figure 5: The complete results of the distribution of [PITH_FULL_IMAGE:figures/full_fig_p035_5.png]
Figure 6
Figure 6. Figure 6: The visualization of the image inverse problems with the base flow matching model of mini-batch optimal [PITH_FULL_IMAGE:figures/full_fig_p037_6.png]
Figure 7
Figure 7. Figure 7: The visualization of the image inverse problems with the base flow matching model of conditional flow matching [PITH_FULL_IMAGE:figures/full_fig_p037_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

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

  1. Optimization and Generation in Aerodynamics Inverse Design

    cs.LG 2026-02 reject novelty 6.0 of 10

    A visual-shape prior reweighted by aerodynamic cost, with a new predictor loss and secant-based covariance guidance, is claimed to reduce drag while preserving design features.

  2. Flow Matching for Efficient and Scalable Data Assimilation

    stat.ML 2025-08 unverdicted novelty 6.0 of 10

    A training-free flow matching data assimilation filter, EnFF, generalizes the bootstrap particle filter and ensemble Kalman filter and reports improved cost-accuracy tradeoffs on high-dimensional benchmarks.

  3. Is Energy Guidance All You Need? Training-Free Norm Injection for Driving World Models

    cs.CV 2026-07 conditional novelty 5.0 of 10

    Sampling-time energy guidance steers a frozen rectified-flow driving world model's ego trajectory to a braking target, but the generated video does not follow under current joint self-attention.

Reference graph

Works this paper leans on

44 extracted references · 34 canonical work pages · cited by 3 Pith papers

  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]

    Is conditional generative modeling all you need for decision-making? In The Eleventh International Conference on Learning Representations, 2023

    Ajay, A., Du, Y., Gupta, A., Tenenbaum, J., Jaakkola, T., and Agrawal, P. Is conditional generative modeling all you need for decision-making? In The Eleventh International Conference on Learning Representations, 2023

    work page 2023

  3. [3]

    Continuous ensemble weather forecasting with diffusion models

    Andrae, M., Landelius, T., Oskarsson, J., and Lindsten, F. Continuous ensemble weather forecasting with diffusion models. In The Thirteenth International Conference on Learning Representations, 2025. URL https://openreview.net/forum?id=ePEZvQNFDW

    work page 2025

  4. [4]

    D-flow: Differentiating through flows for controlled generation

    Ben-Hamu, H., Puny, O., Gat, I., Karrer, B., Singer, U., and Lipman, Y. D-flow: Differentiating through flows for controlled generation. In Proceedings of the 41st International Conference on Machine Learning, 2024

    work page 2024

  5. [5]

    Boys, B., Girolami, M., Pidstrigach, J., Reich, S., Mosca, A., and Akyildiz, O. D. Tweedie moment projected diffusions for inverse problems. In The Twelfth International Conference on Learning Representations, 2023

    work page 2023

  6. [6]

    Chen, R. T. Q. and Lipman, Y. Flow matching on general geometries. In The Twelfth International Conference on Learning Representations, 2024

    work page 2024

  7. [7]

    T., Klasky, M

    Chung, H., Kim, J., Mccann, M. T., Klasky, M. L., and Ye, J. C. Diffusion posterior sampling for general noisy inverse problems. In The Eleventh International Conference on Learning Representations, 2023

    work page 2023

  8. [8]

    and Nichol, A

    Dhariwal, P. and Nichol, A. Diffusion models beat GANs on image synthesis. In The Thirty-Fifth Annual Conference on Neural Information Processing Systems, 2021

    work page 2021

Show all 44 references
  1. [9]

    and Song, Y

    Dou, Z. and Song, Y. Diffusion posterior sampling for linear inverse problem solving: A filtering perspective. In The Twelfth International Conference on Learning Representations, 2024

    2024

  2. [10]

    DPOK : Reinforcement learning for fine-tuning text-to-image diffusion models

    Fan, Y., Watkins, O., Du, Y., Liu, H., Ryu, M., Boutilier, C., Abbeel, P., Ghavamzadeh, M., Lee, K., and Lee, K. DPOK : Reinforcement learning for fine-tuning text-to-image diffusion models. In The Thirty-seventh Annual Conference on Neural Information Processing Systems, 2023

    2023

  3. [11]

    D4rl: Datasets for deep data-driven reinforcement learning

    Fu, J., Kumar, A., Nachum, O., Tucker, G., and Levine, S. D4rl: Datasets for deep data-driven reinforcement learning. arXiv preprint arXiv:2004.07219, 2020

    2004 arXiv

  4. [12]

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

    2024

  5. [13]

    HERO : Human-feedback efficient reinforcement learning for online diffusion model finetuning

    Hiranaka, A., Chen, S.-F., Lai, C.-H., Kim, D., Murata, N., Shibuya, T., Liao, W.-H., Sun, S.-H., and Mitsufuji, Y. HERO : Human-feedback efficient reinforcement learning for online diffusion model finetuning. In The Thirteenth International Conference on Learning Representati...

    2025

  6. [14]

    and Salimans, T

    Ho, J. and Salimans, T. Classifier-free diffusion guidance. In NeurIPS 2021 Workshop on Deep Generative Models and Downstream Applications, 2022

    2021

  7. [15]

    Denoising diffusion probabilistic models

    Ho, J., Jain, A., and Abbeel, P. Denoising diffusion probabilistic models. In The Thirty-Fourth Annual Conference on Neural Information Processing Systems, 2020

    2020

  8. [16]

    B., and Levine, S

    Janner, M., Du, Y., Tenenbaum, J. B., and Levine, S. Planning with diffusion for flexible behavior synthesis. In Proceedings of the 39th International Conference on Machine Learning, 2022

    2022

  9. [17]

    u dke, D., Schwinn, L., and G \

    Kollovieh, M., Lienen, M., L \"u dke, D., Schwinn, L., and G \"u nnemann, S. Flow matching with gaussian process priors for probabilistic time series forecasting. In The Thirteenth International Conference on Learning Representations, 2025. URL https://openreview.net/forum?id=...

    2025

  10. [18]

    Reinforcement learning and control as probabilistic inference: Tutorial and review

    Levine, S. Reinforcement learning and control as probabilistic inference: Tutorial and review. arXiv preprint arXiv:1805.00909, 2018

    2018 arXiv

  11. [19]

    Lipman, Y., Chen, R. T. Q., Ben-Hamu, H., Nickel, M., and Le, M. Flow matching for generative modeling. In The Eleventh International Conference on Learning Representations, 2023

    2023

  12. [20]

    Lipman, Y., Havasi, M., Holderrieth, P., Shaul, N., Le, M., Karrer, B., Chen, R. T. Q., Lopez-Paz, D., Ben-Hamu, H., and Gat, I. Flow matching guide and code. arXiv preprint arXiv:2412.06264, 2024

    2024 arXiv

  13. [21]

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

    Liu, X., Gong, C., and Liu, Q. Flow straight and fast: Learning to generate and transfer data with rectified flow. In The Eleventh International Conference on Learning Representations, 2023 a

    2023

  14. [22]

    FlowGrad : Controlling the output of generative ODEs with gradients

    Liu, X., Wu, L., Zhang, S., Gong, C., Ping, W., and Liu, Q. FlowGrad : Controlling the output of generative ODEs with gradients. In 2023 IEEE / CVF Conference on Computer Vision and Pattern Recognition ( CVPR ) , 2023 b

    2023

  15. [23]

    Contrastive energy prediction for exact energy-guided diffusion sampling in offline reinforcement learning

    Lu, C., Chen, H., Chen, J., Su, H., Li, C., and Zhu, J. Contrastive energy prediction for exact energy-guided diffusion sampling in offline reinforcement learning. In Proceedings of the 40th International Conference on Machine Learning, 2023

    2023

  16. [24]

    What makes a good diffusion planner for decision making? In The Thirteenth International Conference on Learning Representations, 2025

    Lu, H., Han, D., Shen, Y., and Li, D. What makes a good diffusion planner for decision making? In The Thirteenth International Conference on Learning Representations, 2025. URL https://openreview.net/forum?id=7BQkXXM8Fy

    2025

  17. [25]

    S., Boffi, N

    Ma, N., Goldstein, M., Albergo, M. S., Boffi, N. M., Vanden-Eijnden, E., and Xie, S. SiT : Exploring Flow and Diffusion -based Generative Models with Scalable Interpolant Transformers . In The 18th European Conference on Computer Vision, 2024

    2024

  18. [26]

    and Nickel, M

    Mathieu, E. and Nickel, M. Riemannian continuous normalizing flows. In The Thirty-Fourth Annual Conference on Neural Information Processing Systems, 2020

    2020

  19. [27]

    Owen, A. B. Monte Carlo theory, methods and examples. https://artowen.su.domains/mc/, 2013

    2013

  20. [28]

    Masked autoregressive flow for density estimation

    Papamakarios, G., Pavlakou, T., and Murray, I. Masked autoregressive flow for density estimation. In The Thirty-First Annual Conference on Neural Information Processing Systems, 2017

    2017

  21. [29]

    J., Chen, R

    Pokle, A., Muckley, M. J., Chen, R. T. Q., and Karrer, B. Training-free linear image inverses via flows. In The Twelfth International Conference on Learning Representations, 2024

    2024

  22. [30]

    Learning diffusion priors from observations by expectation maximization

    Rozet, F., Andry, G., Lanusse, F., and Louppe, G. Learning diffusion priors from observations by expectation maximization. In The Thirty-Eighth Annual Conference on Neural Information Processing Systems, 2024

    2024

  23. [31]

    Pseudoinverse-guided diffusion models for inverse problems

    Song, J., Vahdat, A., Mardani, M., and Kautz, J. Pseudoinverse-guided diffusion models for inverse problems. In The Eleventh International Conference on Learning Representations, 2023 a

    2023

  24. [32]

    Loss-guided diffusion models for plug-and-play controllable generation

    Song, J., Zhang, Q., Yin, H., Mardani, M., Liu, M.-Y., Kautz, J., Chen, Y., and Vahdat, A. Loss-guided diffusion models for plug-and-play controllable generation. In Proceedings of the 40th International Conference on Machine Learning, 2023 b

    2023

  25. [33]

    and Ermon, S

    Song, Y. and Ermon, S. Generative modeling by estimating gradients of the data distribution. In The Thirty-third Annual Conference on Neural Information Processing Systems, 2019

    2019

  26. [34]

    Sliced score matching: A scalable approach to density and score estimation

    Song, Y., Garg, S., Shi, J., and Ermon, S. Sliced score matching: A scalable approach to density and score estimation. In Proceedings of The 35th Uncertainty in Artificial Intelligence Conference, 2020

    2020

  27. [35]

    P., Kumar, A., Ermon, S., and Poole, B

    Song, Y., Sohl-Dickstein, J., Kingma, D. P., Kumar, A., Ermon, S., and Poole, B. Score-based generative modeling through stochastic differential equations. In The Ninth International Conference on Learning Representations, 2021

    2021

  28. [36]

    Stimper, V., Schölkopf, B., and Hernández-Lobato, J. M. Resampling base distributions of normalizing flows. In Proceedings of the 25th International Conference on Artificial Intelligence and Statistics, 2022

    2022

  29. [37]

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

    Tong, A., Fatras, K., Malkin, N., Huguet, G., Zhang, Y., Rector-Brooks, J., Wolf, G., and Bengio, Y. Improving and generalizing flow-based generative models with minibatch optimal transport. Transactions on Machine Learning Research, March 2024

    2024

  30. [38]

    L., Yim, J., Tischer, D., Baker, D., Broderick, T., Barzilay, R., and Jaakkola, T

    Trippe, B. L., Yim, J., Tischer, D., Baker, D., Broderick, T., Barzilay, R., and Jaakkola, T. Diffusion probabilistic modeling of protein backbones in 3d for the motif-scaffolding problem. In The Eleventh International Conference on Learning Representations, 2023

    2023

  31. [39]

    Training free guided flow-matching with optimal control

    Wang, L., Cheng, C., Liao, Y., Qu, Y., and Liu, G. Training free guided flow-matching with optimal control. In The Thirteenth International Conference on Learning Representations, 2025. URL https://openreview.net/forum?id=61ss5RA1MM

    2025

  32. [40]

    TFG : Unified training-free guidance for diffusion models

    Ye, H., Lin, H., Han, J., Xu, M., Liu, S., Liang, Y., Ma, J., Zou, J., and Ermon, S. TFG : Unified training-free guidance for diffusion models. In The Thirty-Eighth Annual Conference on Neural Information Processing Systems, 2024

    2024

  33. [41]

    Energy-weighted flow matching for offline reinforcement learning

    Zhang, S., Zhang, W., and Gu, Q. Energy-weighted flow matching for offline reinforcement learning. In The Thirteenth International Conference on Learning Representations, 2025. URL https://openreview.net/forum?id=HA0oLUvuGI

    2025

  34. [42]

    Ensemble kalman diffusion guidance: A derivative-free method for inverse problems

    Zheng, H., Chu, W., Wang, A., Kovachki, N., Baptista, R., and Yue, Y. Ensemble kalman diffusion guidance: A derivative-free method for inverse problems. arXiv preprint arXiv:2409.20175v1, 2024

    2024 arXiv

  35. [43]

    Zheng, Q., Le, M., Shaul, N., Lipman, Y., Grover, A., and Chen, R. T. Q. Guided flows for generative modeling and decision making. arXiv preprint arXiv:2311.13443, 2023

    2023 arXiv

  36. [44]

    V., Guntupalli, J

    Zhou, G., Swaminathan, S., Raju, R. V., Guntupalli, J. S., Lehrach, W., Ortiz, J., Dedieu, A., Lázaro-Gredilla, M., and Murphy, K. Diffusion model predictive control. arXiv preprint arXiv:2410.05364, 2024

    2024 arXiv

Pith tools

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