mlsanityData to 7 Oct 2026

Load--Reserve Wasserstein Propagation for Isotropic Diffusion Samplers

arXiv 2603.19670 ↗First posted 20 Mar 2026cs.LGcited by 1

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Many Wasserstein analyses of diffusion samplers control reverse-time propagation by global stability summaries of the learned drift. These summaries can hide radial geometry: equal-height expansive regions of different width can yield different propagation costs. We give a profile-adapted propagation interface for scalar-isotropic reverse-SDE windows with certified learned-drift profiles. A certified lower radial profile is compiled into an affine-tail transportation cost: reflection coupling reduces stability to a one-dimensional slope budget, and Hardy capacity quantifies the load paid before a contractive tail reserve. The compiler yields an adapted cost, contraction rate, and retained tail slope. Score-modeling and solver residuals are treated as forcing inputs and propagate additively in the adapted Wasserstein distance. Quadratic Wasserstein error is reported only at terminal time, using the retained tail slope with tail, moment, or support information. Gaussian-smoothed denoising geometry supplies inverse-radius profiles for uniformly dissipative, bounded-amplitude, and common-covariance mixture windows. Fixed-height examples show that adverse height, even with eventual reserve, does not determine the certificate; barrier examples show that the load dependence is structural.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 27 of its 43 references that are in the map. First posted on arXiv 20 Mar 2026.

Builds on

27 of 43 references.

Closest ancestors · 23

Same direction, newest first

  1. 2026High-accuracy sampling for diffusion models and log-concave distributions1 Feb
  2. 2026On Forgetting and Stability of Score-based Generative models29 Jan
  3. 2025Non-asymptotic error bounds for probability flow ODEs under weak log-concavity20 Oct
  4. 2025Convergence of Deterministic and Stochastic Diffusion-Model Samplers: A Simple Analysis in Wasserstein Distance5 Aug
  5. 2025Wasserstein Convergence of Score-based Generative Models under Semiconvexity and Discontinuous Gradients6 May
  6. 2025Advancing Wasserstein Convergence Analysis of Score-Based Models: Insights from Discretization and Second-Order Acceleration7 Feb
  7. 2025Low-dimensional adaptation of diffusion models: Convergence in total variation22 Jan
  8. 2025Beyond Log-Concavity and Score Regularity: Improved Convergence Bounds for Score-Based Generative Models in W2-distance4 Jan
  9. 2024Wasserstein Bounds for generative diffusion models with Gaussian tail targets15 Dec
  10. 2024Provable Acceleration for Diffusion Models under Minimal Assumptions30 Oct
  11. 2024Unified Convergence Analysis for Score-Based Diffusion Models with Deterministic Samplers18 Oct
  12. 2024O(d/T) Convergence Theory for Diffusion Probabilistic Models under Minimal Assumptions27 Sep
11 more
  1. 2024Score-based generative models break the curse of dimensionality in learning a family of sub-Gaussian probability distributions12 Feb
  2. 2024Convergence Analysis for General Probability Flow ODEs of Diffusion Models in Wasserstein Distances31 Jan
  3. 2023Wasserstein Convergence Guarantees for a General Class of Score-Based Generative Models18 Nov
  4. 2023KL Convergence Guarantees for Score diffusion models under minimal data assumptions23 Aug
  5. 2023Nearly d-Linear Convergence Bounds for Diffusion Models via Stochastic Localization7 Aug
  6. 2023The probability flow ODE is provably fast19 May
  7. 2022Improved Analysis of Score-based Generative Modeling: User-Friendly Bounds under Minimal Smoothness Assumptions3 Nov
  8. 2022Convergence of score-based generative modeling for general data distributions26 Sep
  9. 2022Sampling is as easy as learning the score: theory for diffusion models with minimal data assumptions22 Sep
  10. 2020Score-Based Generative Modeling through Stochastic Differential Equations26 Nov
  11. 2020Denoising Diffusion Probabilistic Models19 Jun

Other references · 4

  1. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  2. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
  3. 2024Optimal score estimation via empirical Bayes smoothingcited by 60
  4. 2018Non-asymptotic bounds for sampling algorithms without log-concavitycited by 35

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

Similar, not cited

Close in wording, no citation either way

  1. 2023Restoration-Degradation Beyond Linear Diffusions: A Non-Asymptotic Analysis For DDIM-Type Samplers
  2. 2026Wasserstein bounds for denoising diffusion probabilistic models via the F"ollmer process
  3. 2025Assessing the Quality of Denoising Diffusion Models in Wasserstein Distance: Noisy Score and Optimal Bounds
  4. 2026Score Accuracy Along the Forward Diffusion Does Not Certify Numerical Stability in Diffusion Sampling
  5. 2026Wasserstein Convergence of ODE-Based Samplers in Decentralized Diffusion Model via Velocity Field Decomposition
  6. 2026Direct Intermediate Initialization for Tilted Diffusion Samplers
  7. 2026On the Robustness of Distribution Support under Diffusion Guidance
  8. 2026Symmetric Equilibrium Propagation for Thermodynamic Diffusion Training