Pith. sign in

REVIEW 14 cited by

Gotta Go Fast When Generating Data with Score-Based Models

Not yet reviewed by Pith; the record is open.

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

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2105.14080 v1 pith:FYZJ533R submitted 2021-05-28 cs.LG cs.CVmath.OCstat.ML

Alexia Jolicoeur-Martineau , Ke Li , Rémi Piché-Taillefer , Tal Kachman , Ioannis Mitliagkas This is my paper · ORCID
classification cs.LGcs.CVmath.OCstat.ML
keywords datasolvermodelssamplesscore-basedstepapproachesdiffusion
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Score-based (denoising diffusion) generative models have recently gained a lot of success in generating realistic and diverse data. These approaches define a forward diffusion process for transforming data to noise and generate data by reversing it (thereby going from noise to data). Unfortunately, current score-based models generate data very slowly due to the sheer number of score network evaluations required by numerical SDE solvers. In this work, we aim to accelerate this process by devising a more efficient SDE solver. Existing approaches rely on the Euler-Maruyama (EM) solver, which uses a fixed step size. We found that naively replacing it with other SDE solvers fares poorly - they either result in low-quality samples or become slower than EM. To get around this issue, we carefully devise an SDE solver with adaptive step sizes tailored to score-based generative models piece by piece. Our solver requires only two score function evaluations, rarely rejects samples, and leads to high-quality samples. Our approach generates data 2 to 10 times faster than EM while achieving better or equal sample quality. For high-resolution images, our method leads to significantly higher quality samples than all other methods tested. Our SDE solver has the benefit of requiring no step size tuning.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 14 Pith papers

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

  1. Asymptotic Preservation and Uniform Accuracy of Diffusion and Flow-Matching Samplers

    cs.LG 2026-07 conditional novelty 7.5 of 10

    DDIM (σ-clock Euler) is the unique layer-exact fixed-step sampler; deterministic residual budgets stay O(1) with no log(1/σ_min), while stochastic path-KL scales as Λ²/N from the Itô term alone.

  2. Faster Diffusion Models via Higher-Order Approximation

    cs.LG 2025-06 conditional novelty 7.0 of 10

    A new higher-order ODE sampler for diffusion models is proven to reach ε total-variation accuracy with eO(d^{1+2/K}/ε^{1/K}) iterations under mild assumptions.

  3. Fast Video Generation with Sliding Tile Attention

    cs.CV 2025-02 conditional novelty 7.0 of 10

    Sliding tile attention (STA) replaces full 3D attention in video diffusion transformers with dense tile-local windows, achieving 1.89x training-free and up to 3.53x fine-tuned end-to-end speedups on HunyuanVideo with ...

  4. Amortized Moment Matching for Visual Generation

    cs.LG 2026-07 accept novelty 6.0 of 10

    Amortized Fréchet Distance uses neural nets to match conditional means and covariances, yielding stronger one-step visual generators than explicit FD-loss or multi-step teachers.

  5. Hierarchical Bayesian inference with compositional score modeling for stellar streams

    astro-ph.GA 2026-07 conditional novelty 6.0 of 10

    Combining three stellar streams and the rotation curve with compositional score modeling yields a posterior for the Milky Way potential—mildly oblate inner halo—but the underlying simulator fails a misspecification test.

  6. Scale-Adaptive Generative Flows for Multiscale Scientific Data

    stat.ML 2025-09 conditional novelty 6.0 of 10

    For generative flows on multiscale scientific fields, the noise spectrum should be at least as rough as the data's, and a scale-adaptive schedule can tame the terminal-time stiffness of rougher noise.

  7. Single-Step Consistent Diffusion Samplers

    cs.LG 2025-02 conditional novelty 6.0 of 10

    Consistent diffusion samplers (CDDS and SCDS) generate samples from unnormalized densities in one or two neural network evaluations, instead of the hundreds used by standard diffusion samplers.

  8. Distributional Diffusion Models with Scoring Rules

    cs.LG 2025-02 conditional novelty 6.0 of 10

    Training diffusion models to output a sample from the posterior distribution of clean data given noisy data, via an energy scoring rule, improves few-step generation quality on images and robot trajectories.

  9. Provable diffusion-based posterior sampling for linear inverse problems via DDIM

    cs.LG 2026-07 reject novelty 5.0 of 10

    A SVD-based, coordinate-wise DDIM sampler is claimed to asymptotically sample from the posterior for noisy linear inverse problems, but the proof's posterior identification step does not follow from the stated updates.

  10. Time-Unified Diffusion Policy with Action Discrimination for Robotic Manipulation

    cs.RO 2025-06 conditional novelty 5.0 of 10

    TUDP removes timestep conditioning from diffusion policies and adds an action-discrimination signal to learn a time-unified velocity field, achieving SOTA RLBench success rates (82.6% multi-view, 83.8% single-view) an...

  11. Unleashing High-Quality Image Generation in Diffusion Sampling Using Second-Order Levenberg-Marquardt-Langevin

    cs.CV 2025-05 reject novelty 5.0 of 10

    A training-free 'Levenberg-Marquardt-Langevin' diffusion sampler is claimed to improve image FID, but its update rule collapses to that of the baseline DPM-Solver for the parameter values used in the paper.

  12. A Diffusion Model Translator for Efficient Image-to-Image Translation

    cs.CV 2025-02 conditional novelty 5.0 of 10

    DMT trains a lightweight translator on equally noised source-target pairs at an intermediate diffusion timestep, then denoises with a frozen DDPM, achieving faster image-to-image translation.

  13. Censored Sampling for Topology Design: Guiding Diffusion with Human Preferences

    cs.LG 2025-08 unverdicted novelty 4.0 of 10

    Guiding a pretrained topology-diffusion generator with human-preference reward classifiers is claimed to suppress floating-material and boundary-violation failure modes without retraining the generator.

  14. Efficient Diffusion Models: A Survey

    cs.LG 2025-02 conditional novelty 2.0 of 10

    The paper organizes research on efficient diffusion models into a taxonomy spanning algorithms, systems, and frameworks, and provides a curated reference list.

Pith tools