Pith. sign in

REVIEW 3 major objections 4 minor 59 references

Single-Step Consistent Diffusion Samplers

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

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

Pith's one-line read A self-consistency loss lets one neural network evaluation produce high-fidelity samples from an unnormalized density, matching 128-step diffusion samplers on the benchmarks tested in this paper.

desk verdict A genuine from-scratch consistency sampler for unnormalized densities, with a bootstrap caveat that needs a sharper base-case guarantee. read the letter →

arxiv 2502.07579 v1 pith:M5NABKZA submitted 2025-02-11 cs.LG stat.ML

Pascal Jutras-Dubé , Patrick Pynadath , Ruqi Zhang This is my paper · ORCID
classification cs.LGstat.ML
keywords unnormalizeddensitysamplingsingle-stepdiffusionsamplersconsistencymodelsself-consistencylossprobabilityflowODElog-variancedivergenceoptimalcontrolfor
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 tries to make sampling from unnormalized densities a one-step operation. It introduces consistent diffusion samplers, which learn a mapping from any intermediate diffusion state straight to a final sample; a distilled variant (CDDS) does this from a pretrained diffusion sampler without collecting a dataset of full samples, and a self-consistent variant (SCDS) trains a single model from scratch with no pretrained sampler at all. SCDS conditions its control on step size and enforces that a large jump equals two small jumps through a self-consistency loss, while the smallest step is learned by a log-variance divergence objective that keeps exploration alive. If the construction works as claimed, practitioners could replace hundreds of numerical integrator calls with one pass through a network and still obtain samples competitive with 128-step diffusion samplers on the tested benchmarks.

What carries the argument

The load-bearing object is the step-size-conditioned control $u_\theta(x,t,d)$ that drives both the SDE and the probability-flow ODE; setting $d=T/N$ reproduces ordinary diffusion steps, while $d=T-t$ reproduces single-step consistency jumps. The self-consistency loss $L_{\mathrm{SC}}$ forces a step of size $2d$ to agree with two composed steps of size $d$, using a stop-gradient target to stabilize recursion, and the base case $d=T/N$ is trained with the log-variance divergence so that Brownian exploration discovers modes before shortcuts are learned.

What would settle it

Train SCDS on a high-dimensional target with separated modes, compare its one-step samples with a long-run reference, and also measure the base-case error by comparing $u_\theta(\cdot,\cdot,T/N)$ against a very fine probability-flow ODE integration; if the one-step samples miss modes while the fine integration does not, the self-consistency loss has spread an inaccurate base case into the single-step result.

Watch

Extended reading notes

Core claim

The central claim is that a single model can simultaneously act as a diffusion sampler and as a consistency model: it learns the small-step dynamics of the probability-flow ODE at the base resolution and, through the self-consistency loss, expands that accuracy to arbitrarily large steps, so a full prior-to-target jump is one Euler update. The paper further claims that the same framework can be distilled from an existing diffusion sampler without storing a pre-collected dataset, and that the from-scratch version SCDS requires no pretrained model, only pointwise evaluations of the unnormalized density. In the reported benchmarks, one-step SCDS and CDDS obtain Sinkhorn distances comparable to the 128-step baselines while using 1% of the network evaluations, and SCDS additionally yields an estimate of the normalizing constant $Z$ through its control representation.

Load-bearing premise

The whole construction depends on the smallest step being right: the base-case transition $d=T/N$ is learned only by a log-variance objective, and no theorem guarantees that this small-step control is accurate, so any error there is inherited by the single-step sample.

Editorial extensions

If this is right

  • Single-step sampling becomes a single Euler update from the prior: draw $x_0$, evaluate $u_\theta(x_0,0,T)$, output $x_T$, reducing network evaluations to about 1% of a 128-step diffusion sampler.
  • The same trained SCDS model can also be run with $K$ steps for refinement, so practitioners can trade latency against accuracy without retraining.
  • CDDS distills a pretrained diffusion sampler by reusing intermediate noisy states and partial ODE trajectories, removing the need to pre-store a large sample dataset.
  • SCDS needs no pretrained teacher and no dataset from the target; it works from pointwise evaluations of the unnormalized density, and can estimate the normalizing constant $\log Z$.
  • Because consistency errors accumulate as $O((\Delta t)^p)$ under a Lipschitz consistency function (Theorem 4.1), finer base discretizations provably shrink the gap between the learned single-step map and the true PF-ODE consistency map.

Reading between the lines

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

  • Editorial inference: the same step-size-conditioned consistency trick could be applied to other base objectives besides the log-variance divergence, letting a user match the base-case training to the geometry of a specific target, such as a KL objective for smooth unimodal targets.
  • Editorial inference: the base case $d=T/N$ is the single point of failure; a practical diagnostic would be to measure how well $u_\theta(\cdot,\cdot,T/N)$ matches a fine-grid PF-ODE transition before spending compute on self-consistency training.
  • Editorial inference: if SCDS scales to higher dimensions, single-step amortization could make posterior sampling with expensive likelihoods affordable in Bayesian inverse problems, where each likelihood evaluation is currently paid thousands of times per chain.
  • Editorial inference: a testable extension is to apply SCDS to time-dependent or conditional targets; the control representation already carries $t$ and $d$ as conditioning variables, so conditioning on additional context may transfer without architectural changes.
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. This paper proposes two samplers for unnormalized densities. CDDS adapts consistency distillation to diffusion samplers: using a pretrained control u, it trains a consistency function f_θ by matching f_θ on consecutive PF-ODE states, without pre-collecting a dataset of terminal samples. SCDS removes the pretrained sampler: a single step-size-conditioned control u_θ(x_t,t,d) is trained jointly with a log-variance sampling loss at the base step d=T/N and a self-consistency loss that asks a 2d Euler step to match two d-step transitions of the same network. At inference, SCDS can sample in one step by taking d=T, or refine with K steps. The paper reports Sinkhorn distances on five benchmark targets and logZ errors on a 1600d LGCP, claiming competitive quality at NFE=1 or 2 relative to 128-step baselines.

Significance. If the claims hold, single-step sampling from unnormalized densities with no pretrained teacher and no dataset would be a substantial practical advance, especially with the added normalizing-constant estimator. The experimental design is strong: diverse targets from 2d to 1600d, comparison to PIS/DDS/DIS, and both sample-quality and logZ metrics. The CDDS data-free distillation idea is clean, and the step-size-conditioned control in SCDS is an original mechanism. The main weakness is theoretical: the only formal guarantee, Theorem 4.1, is for CDDS with a fixed control, and SCDS's self-consistency construction has no convergence or accuracy theorem. The empirical results make the claims plausible, but the central from-scratch single-step assertion currently rests on an unexamined base-case premise.

major comments (3)
  1. [§5, Eqs. (10)–(13)] The central from-scratch claim rests on an unproven bootstrap. In the self-consistency loss (Eq. 12), the target x'_{t+2d} is generated by the same network u_θ at step size d with stopgrad, so for d > T/N the loss only asks that one large step agree with two smaller steps of the same learned control. Accuracy relative to the PF ODE (Eq. 8) is anchored exclusively at the base case d = T/N, where u_θ is trained with the log-variance loss LS (Eq. 13). No theorem or numerical diagnostic establishes that this base-case control is an accurate Euler transition of the PF ODE, and Theorem 4.1 does not cover SCDS because it assumes a fixed control u and a consistency function f(·;u), while SCDS trains u_θ(·,·,d) jointly. The empirical results in Table 1 suggest the base case is adequate for the tested tasks (SCDS-128 is close to DIS-128), but the single-step fidelity claim requires either a convergence argument for the recursive halving construction or a direct evaluation of base-case accuracy, such as comparing u_θ(·,·,T/N) Euler steps against a high-resolution PF ODE reference at increasing d.
  2. [Appendix A, proof of Theorem 4.1] The proof equates a zero expectation with pointwise equality. The loss LCD in Eq. (9) is defined as an expectation over sampled states, so LCD(θ,θ;u)=0 implies λ(tn)d(fθ(x̂_{tn+1},tn+1), fθ(x̂_{tn},tn))=0 only almost surely for the sampled points, not for every xtn appearing in the sup of the theorem. Consequently, the recurrence and the conclusion sup_{n,xtn} ||fθ(xtn,tn) − f(xtn,tn;u)|| = O((Δt)^p) do not follow as stated. A correct statement would need either a pointwise loss or an almost-sure/in-probability bound, and the distinction matters because the same gap is what would be needed to extend the argument to SCDS.
  3. [§5, 'Learning the Base Case' and §6, Table 2] The base-case objective LS is the log-variance divergence for the SDE in Eq. (2), but self-consistency is enforced on Euler steps of the PF ODE in Eq. (8). These are different objects: a control can yield a small DLV while having non-negligible deterministic Euler discretization error, particularly for multi-scale targets such as the Funnel distribution. The paper does not quantify this gap, and the logZ estimation claim inherits the same issue: the estimator is motivated through DKL in Eq. (6), yet the trained control is optimized for DLV (Eq. 13). A bias/variance analysis of the logZ estimator, or an experiment comparing DKL- and DLV-trained controls, would clarify whether the reported logZ errors in Table 2 reflect the method or an artifact of objective mismatch.
minor comments (4)
  1. [Algorithm 2] The parameter update line reads 'θ ← ∇θ (λS(t)LS + λSC(t)LSC)', which is missing the learning rate and the subtraction; it should read θ ← θ − η ∇θ(λS(t)LS + λSC(t)LSC), as in Algorithm 1.
  2. [Eq. (8)] The notation in Eq. (8) is unclear: the term '1/2 σ(t), u(xt,t)' uses an undefined σ and an apparent comma; it should be aligned with Eq. (2), presumably as 1/2 g(t)u_θ(xt,t) or the corresponding PF-ODE drift, with all coefficients defined.
  3. [Tables 1 and 2] The reported Sinkhorn distances and logZ errors appear to come from single runs without standard errors or multiple seeds; given the stochasticity of training and the small differences between SCDS-1 and CDDS-1 on several tasks, the authors should state whether the results are representative or provide variance estimates.
  4. [Abstract and §6] The claim of 'less than 1% of the network evaluations' refers only to inference-time NFE; the training cost of SCDS, which simulates SDE trajectories and adds three network evaluations per iteration, should be stated explicitly to avoid a misleading comparison with traditional samplers.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SCDS's self-consistency loss is a self-distillation bootstrap anchored by the base-case log-variance objective; single-step accuracy is trained, not fitted, and no load-bearing self-citation chain is present.

full rationale

The paper's central single-step claim is not equivalent to its inputs by construction. The base case d = T/N is trained with the log-variance divergence (Eq. 13), which is an external objective evaluated against the unnormalized density through R, S, and B in Eq. 5. The self-consistency loss (Eq. 12) is a recursive bootstrap: large-step controls are trained to match compositions of the same network's smaller-step controls with stopgrad. This is a standard self-distillation/progressive-distillation pattern, and it propagates rather than manufactures base-case accuracy. The only formal guarantee (Theorem 4.1) is stated for CDDS with a fixed control u and a fixed reference consistency function f(.;u), and the paper does not invoke it for SCDS; the absence of a guarantee for SCDS's bootstrapped recursion is a correctness gap, not a circular reduction. There are no load-bearing self-citations by the present authors, no fitted parameter being relabeled as a prediction, and no uniqueness theorem imported from prior work. The experimental comparisons against DIS, DDS, and PIS provide external benchmarks, and the claimed single-step behavior is trained rather than derived from the objective itself. Therefore the derivation chain is self-contained apart from the noted missing-guarantee concern, which warrants no circularity score.

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

The methods assume standard probability-flow and stochastic-control results from the cited literature, plus a base-case accuracy premise for SCDS that is not proven. No new physical or mathematical entities are introduced.

free parameters (2)
  • Loss weight schedules lambda_S(t), lambda_SC(t) = Not specified
    Algorithm 2 uses lambda_S(t)LS + lambda_SC(t)LSC but the schedules are not given in the paper; these weights control the balance between exploration and consistency and are likely tuned per task.
  • Initial distribution truncation quartile (1e-4) = 1e-4
    The initial distribution is N(0,I) with truncation at a quartile of 1e-4; this is a hand-chosen detail that affects the support of the starting noise.
assumptions (5)
  • standard math The PF ODE (Eq. 8) has the same marginal distributions as the SDE (Eq. 2) for a given control u.
    Used in both CDDS and SCDS to justify training with deterministic trajectories instead of the stochastic SDE. This is the standard probability-flow ODE result of Song et al. 2021b.
  • domain assumption The log-variance divergence DLV (Eq. 7) is a valid, optimizable objective whose minimizer yields the optimal control u* = g grad log p.
    Adopted from Richter and Berner 2024 and used as the base sampling loss LS for SCDS (Eq. 13).
  • domain assumption A pretrained diffusion sampler (DIS) with control u is available and its PF ODE trajectories are reliable teacher signals for CDDS.
    CDDS initializes from DIS weights and distills the PF ODE of DIS; the quality of the teacher bounds the quality of the distilled sampler.
  • ad hoc to paper The consistency function f_theta satisfies a Lipschitz condition and the boundary condition f_theta(x_T,T) = x_T.
    These assumptions are stated in Theorem 4.1 and are needed for the error bound O((Delta t)^p).
  • domain assumption Euler discretization of the PF ODE (Eqs. 10-11) is an adequate integration scheme for the self-consistency loss; the control vector field is smooth enough for local errors to be controlled.
    Used to define the shortcut and two-step targets in SCDS; if the Euler steps are very inaccurate for the chosen d, the consistency loss will enforce consistency to an inaccurate base.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Single-Step Consistent Diffusion Samplers." pith.science (2026). https://pith.science/paper/M5NABKZA

@misc{pith2026250207579,
  author       = {Pith},
  title        = {Pith review of: Single-Step Consistent Diffusion Samplers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/M5NABKZA}},
  note         = {Machine review of arXiv:2502.07579}
}
read the original abstract

Sampling from unnormalized target distributions is a fundamental yet challenging task in machine learning and statistics. Existing sampling algorithms typically require many iterative steps to produce high-quality samples, leading to high computational costs that limit their practicality in time-sensitive or resource-constrained settings. In this work, we introduce consistent diffusion samplers, a new class of samplers designed to generate high-fidelity samples in a single step. We first develop a distillation algorithm to train a consistent diffusion sampler from a pretrained diffusion model without pre-collecting large datasets of samples. Our algorithm leverages incomplete sampling trajectories and noisy intermediate states directly from the diffusion process. We further propose a method to train a consistent diffusion sampler from scratch, fully amortizing exploration by training a single model that both performs diffusion sampling and skips intermediate steps using a self-consistency loss. Through extensive experiments on a variety of unnormalized distributions, we show that our approach yields high-fidelity samples using less than 1% of the network evaluations required by traditional diffusion samplers.

Figures

Figures reproduced from arXiv: 2502.07579 by the authors.

Figure 1
Figure 1. Consistency distilled diffusion samplers learn to map consecutive intermediate states (black and gray dots) along partial ODE trajectories (green curve) directly to the terminal state. If the loss in Eq. 9 is driven to zero, the learned consistency function can approximate the true mapping arbitrarily well, provided the step size of the ODE solver is sufficiently small. We formally state this in Theorem 4.1. Theorem… view at source ↗
Figure 2
Figure 2. Graphical illustration of the training procedure for SCDS over the path space. First, the SDE trajectory (white) is simulated to compute the sampling loss LS. Next, a timestep t and a step size d are randomly sampled. From xt on the simulated SDE trajectory, we execute two consecutive steps of size d (red) along the PF-ODE trajectory (pink), obtaining the target x ′ t+2d. Finally, the shortcut step of size d (orange… view at source ↗
Figure 3
Figure 3. Visualization of the GMM and MW54 tasks. CDDS and SCDS recover all modes in just a single sampling step. network to condition on the stepsize d (Tancik et al., 2020). When ground-truth samples are available, we measure per￾formance via the Sinkhorn distance (Cuturi, 2013) between generated samples and samples from the target distribution. For the LGCP task, we report the relative error of the esti￾mated normalizing … view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Comparison of Sinkhorn distance for a range of NFEs between the proposed consistency samplers (CDDS, SCDS) and diffusion￾based samplers (PIS, DDS, DIS). For most targets, CDDS and SCDS show competitive Sinkhorn values with baselines with much lower NFEs [PITH_FULL_IMA…
Figure 5
Figure 5. Figure 5 [PITH_FULL_IMAGE:figures/full_fig_p015_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 47 canonical work pages

  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]

    S., Kanwar, G., and Shanahan, P

    Albergo, M. S., Kanwar, G., and Shanahan, P. E. Flow-based generative models for markov chain monte carlo in lattice field theory. Physical Review D, 100 0 (3): 0 034515, 2019

    work page 2019

  3. [3]

    Anderson, B. D. O. Reverse-time diffusion equation models. Stochastic Processes and their Applications, 12: 0 313--326, 1982

    work page 1982

  4. [4]

    Andrieu, C., de Freitas, N., Doucet, A., and Jordan, M. I. An introduction to mcmc for machine learning. Machine Learning, 50 0 (1-2): 0 5--43, 2003

    work page 2003

  5. [5]

    Annealed flow transport monte carlo

    Arbel, M., Matthews, A., and Doucet, A. Annealed flow transport monte carlo. In International Conference on Machine Learning, pp.\ 318--330. PMLR, 2021

    work page 2021

  6. [6]

    Flow network based generative models for non-iterative diverse candidate generation

    Bengio, E., Jain, M., Korablyov, M., Precup, D., and Bengio, Y. Flow network based generative models for non-iterative diverse candidate generation. In Advances in Neural Information Processing Systems, 2021

    work page 2021

  7. [7]

    J., Tiwari, M., and Bengio, E

    Bengio, Y., Lahlou, S., Deleu, T., Hu, E. J., Tiwari, M., and Bengio, E. Gflownet foundations. Journal of Machine Learning Research, 24 0 (210): 0 1--55, 2023

    2023

  8. [8]

    An optimal control perspective on diffusion-based generative modeling

    Berner, J., Richter, L., and Ullrich, K. An optimal control perspective on diffusion-based generative modeling. Transactions on Machine Learning Research, 2024. ISSN 2835-8856

    work page 2024

Show all 59 references
  1. [9]

    P., Gelman, A., Jones, G

    Brooks, S. P., Gelman, A., Jones, G. L., and Meng, X.-L. Handbook of markov chain monte carlo: Hardcover. CHANCE, 25: 0 53--55, 2012

    2012

  2. [10]

    Sequential controlled langevin diffusions

    Chen, J., Richter, L., Berner, J., Blessing, D., Neumann, G., and Anandkumar, A. Sequential controlled langevin diffusions. In Proceedings of the International Conference on Learning Representations (ICLR), 2025

    2025

  3. [11]

    Stochastic gradient hamiltonian monte carlo

    Chen, T., Fox, E., and Guestrin, C. Stochastic gradient hamiltonian monte carlo. In International Conference on Machine Learning, pp.\ 1683--1691. PMLR, 2014

    2014

  4. [12]

    Sinkhorn distances: Lightspeed computation of optimal transport

    Cuturi, M. Sinkhorn distances: Lightspeed computation of optimal transport. In Advances in Neural Information Processing Systems, pp.\ 2292--2300, 2013

    2013

  5. [13]

    J., and Roweth, D

    Duane, S., Kennedy, A., Pendleton, B. J., and Roweth, D. Hybrid monte carlo. Physics Letters B, 195 0 (2): 0 216--222, 1987. ISSN 0370-2693

    1987

  6. [14]

    One step diffusion via shortcut models

    Frans, K., Hafner, D., Levine, S., and Abbeel, P. One step diffusion via shortcut models. In Proceedings of the International Conference on Learning Representations. ICLR, 2025

    2025

  7. [15]

    and Smit, B

    Frenkel, D. and Smit, B. Understanding Molecular Simulation: From Algorithms to Applications. Academic Press, Amsterdam, The Netherlands, 2002. ISBN 978-0-12-267351-1

    2002

  8. [16]

    and Domke, J

    Geffner, T. and Domke, J. Langevin diffusion variational inference. In International Conference on Artificial Intelligence and Statistics, pp.\ 576--593. PMLR, 2023

    2023

  9. [17]

    and Geman, D

    Geman, S. and Geman, D. Stochastic relaxation, gibbs distributions, and the bayesian restoration of images. IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI-6 0 (6): 0 721--741, 1984

    1984

  10. [18]

    Gershman, S. J. and Goodman, N. D. Amortized inference in probabilistic reasoning. In Proceedings of the Annual Meeting of the Cognitive Science Society. Stanford University, 2014

    2014

  11. [19]

    Hastings, W. K. Monte carlo sampling methods using markov chains and their applications. Biometrika, 57 0 (1): 0 97--109, 1970

    1970

  12. [20]

    Hern \'a ndez-Lobato, J. M. and Adams, R. Probabilistic backpropagation for scalable learning of bayesian neural networks. In International Conference on Machine Learning, pp.\ 1861--1869. PMLR, 2015

    2015

  13. [21]

    Denoising diffusion probabilistic models

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

    2020

  14. [22]

    Stochastic optimal control for collective variable free sampling of molecular transition paths

    Holdijk, L., Du, Y., Hooft, F., Jaini, P., Ensing, B., and Welling, M. Stochastic optimal control for collective variable free sampling of molecular transition paths. Advances in Neural Information Processing Systems, 36, 2024

    2024

  15. [23]

    Hollingsworth, S. A. and Dror, R. O. Molecular dynamics simulation for all. Neuron, 99 0 (6): 0 1129--1143, 2018. ISSN 0896-6273

    2018

  16. [24]

    Gotta go fast when generating data with score-based models

    Jolicoeur-Martineau, A., Li, K., Pich \'e -Taillefer, R., Kachman, T., and Mitliagkas, I. Gotta go fast when generating data with score-based models. arXiv preprint arXiv:2105.14080, 2021

    2021 arXiv

  17. [25]

    Elucidating the design space of diffusion-based generative models

    Karras, T., Aittala, M., Aila, T., and Laine, S. Elucidating the design space of diffusion-based generative models. In Advances in Neural Information Processing Systems, 2022

    2022

  18. [26]

    E., and Liu, Q

    Li, Y., Turner, R. E., and Liu, Q. Approximate inference with amortised mcmc. arXiv preprint arXiv:1702.08343, 2017

    2017 arXiv

  19. [27]

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

    2023

  20. [28]

    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 International Conference on Learning Representations, 2023

    2023

  21. [29]

    and Song, Y

    Lu, C. and Song, Y. Simplifying, stabilizing and scaling continuous-time consistency models. In International Conference on Learning Representations, 2025

    2025

  22. [30]

    MacKay, D. J. Information Theory, Inference and Learning Algorithms. Cambridge University Press, 2003

    2003

  23. [31]

    On distillation of guided diffusion models

    Meng, C., Rombach, R., Gao, R., Kingma, D., Ermon, S., Ho, J., and Salimans, T. On distillation of guided diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023

    2023

  24. [32]

    W., Rosenbluth, M

    Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H., and Teller, E. Equation of state calculations by fast computing machines. The Journal of Chemical Physics, 21 0 (6): 0 1087--1092, 1953

    1953

  25. [33]

    Neal, R. M. Bayesian learning for neural networks. 1995

    1995

  26. [34]

    Neal, R. M. Annealed importance sampling. Statistics and Computing, 11: 0 125--139, 2001

    2001

  27. [35]

    Neal, R. M. Slice sampling. The annals of statistics, 31 0 (3): 0 705--767, 2003

    2003

  28. [36]

    Boltzmann generators: Sampling equilibrium states of many-body systems with deep learning

    No \'e , F., K \"o hler, J., and Wu, H. Boltzmann generators: Sampling equilibrium states of many-body systems with deep learning. Science, 365, 2019. URL https://api.semanticscholar.org/CorpusID:54458652

    2019

  29. [37]

    Correlation functions and computer simulations

    Parisi, G. Correlation functions and computer simulations. Nuclear Physics B, 180 0 (3): 0 378--384, 1981

    1981

  30. [38]

    J., Bortoli, V

    Phillips, A., Dau, H.-D., Hutchinson, M. J., Bortoli, V. D., Deligiannidis, G., and Doucet, A. Particle denoising diffusion sampler, 2024

    2024

  31. [39]

    and Berner, J

    Richter, L. and Berner, J. Improved sampling via learned diffusions. In International Conference on Learning Representations, 2024

    2024

  32. [40]

    Robert, C. P. Convergence control methods for markov chain monte carlo algorithms. Statistical Science, 10 0 (3): 0 231--253, 1995

    1995

  33. [41]

    J., Doll, J

    Rossky, P. J., Doll, J. D., and Friedman, H. L. Brownian Dynamics as Smart Monte Carlo Simulation . The Journal of Chemical Physics, 69 0 (10): 0 4628--4633, 11 1978

    1978

  34. [42]

    and Ho, J

    Salimans, T. and Ho, J. Progressive distillation for fast sampling of diffusion models. In International Conference on Learning Representations, 2022

    2022

  35. [43]

    Über die umkehrung der naturgesetze

    Schrödinger, E. Über die umkehrung der naturgesetze. Sitzungsberichte der Preussischen Akademie der Wissenschaften Berlin, Physikalisch-Mathematische Klasse, pp.\ 144--153, 1931

    1931

  36. [44]

    Sur la théorie relativiste de l'Électron et l'interprétation de la mécanique quantique

    Schrödinger, E. Sur la théorie relativiste de l'Électron et l'interprétation de la mécanique quantique. Annales de l'Institut Henri Poincaré, 2: 0 269--310, 1932

    1932

  37. [45]

    Deep unsupervised learning using nonequilibrium thermodynamics

    Sohl-Dickstein, J., Weiss, E., Maheswaranathan, N., and Ganguli, S. Deep unsupervised learning using nonequilibrium thermodynamics. In International Conference on Machine Learning, pp.\ 2256--2265. PMLR, 2015

    2015

  38. [46]

    Denoising diffusion implicit models

    Song, J., Meng, C., and Ermon, S. Denoising diffusion implicit models. In International Conference on Learning Representations, 2021 a

    2021

  39. [47]

    and Dhariwal, P

    Song, Y. and Dhariwal, P. Improved techniques for training consistency models. arXiv preprint arXiv:2310.14189, 2023

    2023 arXiv

  40. [48]

    and Ermon, S

    Song, Y. and Ermon, S. Generative modeling by estimating gradients of the data distribution. Advances in Neural Information Processing Systems, 32, 2019

    2019

  41. [49]

    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 International Conference on Learning Representations, 2021 b

    2021

  42. [50]

    Consistency models

    Song, Y., Dhariwal, P., Chen, M., and Sutskever, I. Consistency models. arXiv preprint arXiv:2303.01469, 2023

    2023 arXiv

  43. [51]

    Fourier features let networks learn high frequency functions in low dimensional domains

    Tancik, M., Srinivasan, P., Mildenhall, B., Fridovich-Keil, S., Raghavan, N., Singhal, U., Ramamoorthi, R., Barron, J., and Ng, R. Fourier features let networks learn high frequency functions in low dimensional domains. Advances in neural information processing systems, 33: 0 ...

    2020

  44. [52]

    S., and Doucet, A

    Vargas, F., Grathwohl, W. S., and Doucet, A. Denoising diffusion samplers. In International Conference on Learning Representations, 2023

    2023

  45. [53]

    and Teh, Y

    Welling, M. and Teh, Y. W. Bayesian learning via stochastic gradient langevin dynamics. In International Conference on Machine Learning, pp.\ 681--688, Madison, WI, USA, 2011. Omnipress. ISBN 9781450306195

    2011

  46. [54]

    Solving statistical mechanics using variational autoregressive networks

    Wu, D., Wang, L., and Zhang, P. Solving statistical mechanics using variational autoregressive networks. Physical Review Letters, 122 0 (8): 0 080602, 2019

    2019

  47. [55]

    Stochastic normalizing flows

    Wu, H., K \"o hler, J., and No \'e , F. Stochastic normalizing flows. Advances in Neural Information Processing Systems, 33: 0 5933--5944, 2020

    2020

  48. [56]

    Zhang, D., Chen, R. T. Q., Liu, C.-H., Courville, A., and Bengio, Y. Diffusion generative flow samplers: Improving learning signals through partial trajectory optimization. In International Conference on Learning Representations, 2024

    2024

  49. [57]

    and Chen, Y

    Zhang, Q. and Chen, Y. Path integral sampler: A stochastic control approach for sampling. In International Conference on Learning Representations, 2022

    2022

  50. [58]

    F., and De Sa, C

    Zhang, R., Cooper, A. F., and De Sa, C. Amagold: Amortized metropolis adjustment for efficient stochastic gradient mcmc. In International Conference on Artificial Intelligence and Statistics, pp.\ 2142--2152. PMLR, 2020 a

    2020

  51. [59]

    Zhang, R., Li, C., Zhang, J., Chen, C., and Wilson, A. G. Cyclical stochastic gradient mcmc for bayesian deep learning. International Conference on Learning Representations, 2020 b

    2020

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.