mlsanityData to 7 Oct 2026

Taming Score-Based Denoisers in ADMM: A Convergent Plug-and-Play Framework

arXiv 2603.10281 ↗First posted 10 Mar 2026cs.LG cs.AI cs.CVcited by 1

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

While score-based generative models have emerged as powerful priors for solving inverse problems, directly integrating them into optimization algorithms such as ADMM remains nontrivial. Two central challenges arise: i) the mismatch between the noisy data manifolds used to train the score functions and the geometry of ADMM iterates, especially due to the influence of dual variables, and ii) the lack of convergence understanding when ADMM is equipped with score-based denoisers. To address the manifold mismatch issue, we propose ADMM plug-and-play (ADMM-PnP) with the AC-DC denoiser, a new framework that embeds a three-stage denoiser into ADMM: (1) auto-correction (AC) via additive Gaussian noise, (2) directional correction (DC) using conditional Langevin dynamics, and (3) score-based denoising. In terms of convergence, we establish two results: first, under proper denoiser parameters, each ADMM iteration is a weakly nonexpansive operator, ensuring high-probability fixed-point textitball convergence using a constant step size; second, under more relaxed conditions, the AC-DC denoiser is a bounded denoiser, which leads to convergence under an adaptive step size schedule. Experiments on a range of inverse problems demonstrate that our method consistently improves solution quality over a variety of baselines.

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 51 of its 89 references that are in the map. First posted on arXiv 10 Mar 2026.

Builds on

51 of 89 references.

Closest ancestors · 5

Same direction, newest first

  1. 2023Manifold Preserving Guided Diffusion28 Nov
  2. 2022Diffusion models as plug-and-play priors17 Jun
  3. 2022Elucidating the Design Space of Diffusion-Based Generative Models1 Jun
  4. 2021Tackling the Generative Learning Trilemma with Denoising Diffusion GANs15 Dec
  5. 2020Score-Based Generative Modeling through Stochastic Differential Equations26 Nov

Other references · 46

  1. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  2. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  3. 2021Convex optimizationcited by 2,479
  4. 2021SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equationscited by 1,903
  5. 2022Diffusion Posterior Sampling for General Noisy Inverse Problemscited by 1,375
  6. 2017Deep Image Priorcited by 1,097
  7. 2022Denoising Diffusion Restoration Modelscited by 951
  8. 2022Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Modelcited by 574
  9. 2021Solving Inverse Problems in Medical Imaging with Score-Based Generative Modelscited by 514
  10. 2022Improving Diffusion Models for Inverse Problems using Manifold Constraintscited by 501
  11. 2022Diffusion Models for Adversarial Purificationcited by 469
  12. 2020Plug-and-Play Image Restoration with Deep Denoiser Priorcited by 467
34 more
  1. 2016Deep Convolutional Neural Network for Inverse Problems in Imagingcited by 431
  2. 2017Compressed Sensing using Generative Modelscited by 337
  3. 2023Denoising Diffusion Models for Plug-and-Play Image Restorationcited by 335
  4. 2016Plug-and-Play ADMM for Image Restoration: Fixed Point Convergence and Applicationscited by 244
  5. 2023A Variational Perspective on Solving Inverse Problems with Diffusion Modelscited by 217
  6. 2019Plug-and-Play Methods Provably Converge with Properly Trained Denoiserscited by 197
  7. 2021SNIPS: Solving Noisy Inverse Problems Stochasticallycited by 191
  8. 2023Solving Inverse Problems with Latent Diffusion Models via Hard Data Consistencycited by 180
  9. 2024Improving Diffusion Inverse Problem Solving with Decoupled Noise Annealingcited by 157
  10. 2017User-friendly guarantees for the Langevin Monte Carlo with inaccurate gradientcited by 116
  11. 2021Gradient Step Denoiser for convergent Plug-and-Playcited by 107
  12. 2024Principled Probabilistic Imaging using Diffusion Models as Plug-and-Play Priorscited by 93
  13. 2022Proximal Denoiser for Convergent Plug-and-Play Optimization with Nonconvex Regularizationcited by 79
  14. 2019Plug and play methods for magnetic resonance imaging (long version)cited by 75
  15. 2018An Online Plug-and-Play Algorithm for Regularized Image Reconstructioncited by 70
  16. 2014From Denoising to Compressed Sensingcited by 67
  17. 2024DMPlug: A Plug-in Method for Solving Inverse Problems with Diffusion Modelscited by 60
  18. 2017Plug-and-Play Unplugged: Optimization Free Reconstruction using Consensus Equilibriumcited by 58
  19. 2018Solving Linear Inverse Problems Using GAN Priors: An Algorithm with Provable Guaranteescited by 58
  20. 2021Explore Image Deblurring via Blur Kernel Spacecited by 58
  21. 2023Provable Probabilistic Imaging using Score-Based Generative Priorscited by 51
  22. 2021Plug-and-Play Algorithms for Video Snapshot Compressive Imagingcited by 42
  23. 2024Improving Diffusion Models for Inverse Problems Using Optimal Posterior Covariancecited by 42
  24. 2013Tensor-based formulation and nuclear norm regularization for multi-energy computed tomographycited by 22
  25. 2020TFPnP: Tuning-free Plug-and-Play Proximal Algorithm with Applications to Inverse Imaging Problemscited by 20
  26. 2024Plug-and-Play image restoration with Stochastic deNOising REgularizationcited by 17
  27. 2019On Stationary-Point Hitting Time and Ergodicity of Stochastic Gradient Langevin Dynamicscited by 14
  28. 2024Diffusion State-Guided Projected Gradient for Inverse Problemscited by 14
  29. 2018Performance Analysis of Plug-and-Play ADMM: A Graph Signal Processing Perspectivecited by 13
  30. 2021A General Derivative Identity for the Conditional Mean Estimator in Gaussian Noise and Some Applicationscited by 7
  31. 2023Robust Physics-based Deep MRI Reconstruction Via Diffusion Purificationcited by 4
  32. 2025Learning Cocoercive Conservative Denoisers via Helmholtz Decomposition for Poisson Inverse Problemscited by 4
  33. 2025Radio Map Estimation via Latent Domain Plug-and-Play Denoisingcited by 3
  34. 2015Tight Global Linear Convergence Rate Bounds for Douglas-Rachford Splittingcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Convergent Plug-and-Play Image Restoration with Annealed Noise Levelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Optimal Denoising in Score-Based Generative Models: The Role of Data Regularity
  2. 2021On tuning consistent annealed sampling for denoising score matching
  3. 2026The Loss Floor of Denoising Score Matching: Fisher Geometry from Schr"odinger Bridges
  4. 2024Discrete vs. Continuous Trade-offs for Generative Models
  5. 2022Convergence of score-based generative modeling for general data distributions
  6. 2023Solving Inverse Problems with Score-Based Generative Priors learned from Noisy Data
  7. 2026Transformers Learn the Optimal DDPM Denoiser for Multi-Token GMMs
  8. 2025Denoising Score Distillation: From Noisy Diffusion Pretraining to One-Step High-Quality Generation