mlsanityData to 14 Oct 2026

Is Noise Conditioning Necessary for Denoising Generative Models?

arXiv 2502.13129 ↗First posted 18 Feb 2025cs.CVcited by 44

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

It is widely believed that noise conditioning is indispensable for denoising diffusion models to work successfully. This work challenges this belief. Motivated by research on blind image denoising, we investigate a variety of denoising-based generative models in the absence of noise conditioning. To our surprise, most models exhibit graceful degradation, and in some cases, they even perform better without noise conditioning. We provide a theoretical analysis of the error caused by removing noise conditioning and demonstrate that our analysis aligns with empirical observations. We further introduce a noise-unconditional model that achieves a competitive FID of 2.23 on CIFAR-10, significantly narrowing the gap to leading noise-conditional models. We hope our findings will inspire the community to revisit the foundations and formulations of denoising generative models.

Key facts. Cited by 44 papers in the map (data through 14 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 24 of its 42 references that are in the map. First posted on arXiv 18 Feb 2025.

Builds on

24 of 42 references.

Closest ancestors · 9

Same direction, newest first

  1. 2024Consistency Models Made Easy20 Jun
  2. 2023Improved Techniques for Training Consistency Models22 Oct
  3. 2023Consistency Models2 Mar
  4. 2022Elucidating the Design Space of Diffusion-Based Generative Models1 Jun
  5. 2021Diffusion Models Beat GANs on Image Synthesis11 May
  6. 2021Improved Denoising Diffusion Probabilistic Models18 Feb
  7. 2020Score-Based Generative Modeling through Stochastic Differential Equations26 Nov
  8. 2020Denoising Diffusion Implicit Models6 Oct
  9. 2020Denoising Diffusion Probabilistic Models19 Jun

Other references · 15

  1. 2017Attention Is All You Needcited by 60,361
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  3. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
  4. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 6,013
  5. 2022Flow Matching for Generative Modelingcited by 5,702
  6. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,101
  7. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,700
  8. 2020Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domainscited by 2,200
  9. 2021Alias-Free Generative Adversarial Networkscited by 1,184
  10. 2023Stochastic Interpolants: A Unifying Framework for Flows and Diffusionscited by 939
  11. 2019On the Variance of the Adaptive Learning Rate and Beyondcited by 772
  12. 2019Noise2Self: Blind Denoising by Self-Supervisioncited by 276
3 more
  1. 2018Toward Convolutional Blind Denoising of Real Photographscited by 274
  2. 2021How to Train Your Energy-Based Modelscited by 274
  3. 2022Practical Blind Image Denoising via Swin-Conv-UNet and Data Synthesiscited by 86

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 16; 2026: 28.

44 papers.

Sort:

By area:

  1. 2025Energy Matching: Unifying Flow Matching and Energy-Based Models for Generative Modelingcited by 27
  2. 2026Speed by Simplicity: A Single-Stream Architecture for Fast Audio-Video Generative Foundation Modelcited by 25
  3. 2025Equilibrium Matching: Generative Modeling with Implicit Energy-Based Modelscited by 21
  4. 2025H3DP: Triply-Hierarchical Diffusion Policy for Visuomotor Learningcited by 16
  5. 2025Exploring the Deep Fusion of Large Language Models and Diffusion Transformers for Text-to-Image Synthesiscited by 16
  6. 2025Uniform Discrete Diffusion with Metric Path for Video Generationcited by 12
  7. 2025Generalization of Diffusion Models Arises with a Balanced Representation Spacecited by 11
  8. 2025SuperFlow: Training Flow Matching Models with RL on the Flycited by 10
  9. 2026Blind denoising diffusion models and the blessings of dimensionalitycited by 9
  10. 2025Scaling Sequence-to-Sequence Generative Neural Renderingcited by 8
  11. 2026The Geometry of Noise: Why Diffusion Models Don't Need Noise Conditioningcited by 8
  12. 2025Flow Matching in the Low-Noise Regime: Pathologies and a Contrastive Remedycited by 3
32 more
  1. 2026WIND: Weather Inverse Diffusion for Zero-Shot Atmospheric Modelingcited by 3
  2. 2026Discrete Stochastic Localization for Non-autoregressive Generationcited by 3
  3. 2026Generative Control as Optimization: Time Unconditional Flow Matching for Adaptive and Robust Robotic Controlcited by 3
  4. 2025Neural Approaches to SAT Solving: Design Choices and Interpretabilitycited by 2
  5. 2025Mixture of States: Routing Token-Level Dynamics for Multimodal Generationcited by 2
  6. 2026DVD: Discrete Voxel Diffusion for 3D Generation and Editingcited by 2
  7. 2026What Time Is It? How Data Geometry Makes Time Conditioning Optional for Flow Matchingcited by 2
  8. 2026CrossFlow: One-Step Generation Across Latent and Pixel Spacescited by 2
  9. 2026Beckmann Transport Models: From Autonomous Flows to One-Step Mapscited by 2
  10. 2026SPROUT: A Scalable Diffusion Foundation Model for Multi-Crop Plant Phenotypingcited by 1
  11. 2026Exploring Time Conditioning in Diffusion Generative Models from Disjoint Noisy Data Manifoldscited by 1
  12. 2026NoiseRater: Meta-Learned Noise Valuation for Diffusion Model Trainingcited by 1
  13. 2026Generative Pseudo-Force Fields for Molecular Generationcited by 1
  14. 2026Time-Unconditional Generative Speech Enhancement via Autonomous Rectified Flowcited by 1
  15. 2026Equilibrium Forcing: Adaptive Video Generation Without Noise Conditioningcited by 1
  16. 2025Is Noise Conditioning Necessary? A Unified Theory of Unconditional Graph Diffusion Modelscited by 0
  17. 2025MAD: Manifold Attracted Diffusioncited by 0
  18. 2025A Geometric Unification of Generative AI with Manifold-Probabilistic Projection Modelscited by 0
  19. 2025Mitigating the Noise Shift for Denoising Generative Models via Noise Awareness Guidancecited by 0
  20. 2025DP-TTA: Test-time Adaptation for Transient Electromagnetic Signal Denoising via Dictionary-driven Prior Regularizationcited by 0
  21. 2026Distance Marching for Generative Modelingcited by 0
  22. 2026MealRec: Multi-granularity Sequential Modeling via Hierarchical Diffusion Models for Micro-Video Recommendationcited by 0
  23. 2026Mean Mode Screaming: Mean--Variance Split Residuals for 1000-Layer Diffusion Transformerscited by 0
  24. 2026Discrete Stochastic Localization for Non-autoregressive Generationcited by 0
  25. 2026Score Based Error Correcting Code Decodercited by 0
  26. 2026Test-Time Adaptation in Optical Coherence Tomography Using Trajectory-Aligned Time-Independent Flowcited by 0
  27. 2026On the Redundancy of Timestep Embeddings in Diffusion Modelscited by 0
  28. 2026Does Uniform Discrete Diffusion Need Time?cited by 0
  29. 2026A statistical theory for blind denoising: minimax estimation of the noise level from a single samplecited by 0
  30. 2026RepFlow: Reciprocal Supervision Improves Generation and Representation in Flow Modelscited by 0
  31. 2026Jumping up and down: Denoiser diffusion models for discrete ordinal datacited by 0
  32. 2026PixARG: Pixel-Asynchronous Retrieval-Augmented Diffusion for Fast and Realistic Image Generationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2023Diffusion Model for Generative Image Denoising
  2. 2025Beyond Fixed Horizons: A Theoretical Framework for Adaptive Denoising Diffusions
  3. 2021Noise Estimation for Generative Diffusion Models
  4. 2025Is Your Conditional Diffusion Model Actually Denoising?
  5. 2022Image Embedding for Denoising Generative Models
  6. 2025Improving Diffusion Generative Models via Truncated Karhunen--Lo`eve Expansion
  7. 2025Back to Basics: Let Denoising Generative Models Denoise
  8. 2024Statistical guarantees for denoising reflected diffusion models