mlsanityData to 7 Oct 2026

Training Unbiased Diffusion Models From Biased Dataset

arXiv 2403.01189 ↗First posted 2 Mar 2024cs.LG cs.CVcited by 25

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

With significant advancements in diffusion models, addressing the potential risks of dataset bias becomes increasingly important. Since generated outputs directly suffer from dataset bias, mitigating latent bias becomes a key factor in improving sample quality and proportion. This paper proposes time-dependent importance reweighting to mitigate the bias for the diffusion models. We demonstrate that the time-dependent density ratio becomes more precise than previous approaches, thereby minimizing error propagation in generative learning. While directly applying it to score-matching is intractable, we discover that using the time-dependent density ratio both for reweighting and score correction can lead to a tractable form of the objective function to regenerate the unbiased data density. Furthermore, we theoretically establish a connection with traditional score-matching, and we demonstrate its convergence to an unbiased distribution. The experimental evidence supports the usefulness of the proposed method, which outperforms baselines including time-independent importance reweighting on CIFAR-10, CIFAR-100, FFHQ, and CelebA with various bias settings. Our code is available at https://github.com/alsdudrla10/TIW-DSM.

Key facts. Cited by 25 papers in the map (data through 7 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 54 of its 81 references that are in the map. First posted on arXiv 2 Mar 2024.

Builds on

54 of 81 references.

Closest ancestors · 21

Same direction, newest first

  1. 2024Label-Noise Robust Diffusion Models27 Feb
  2. 2023Denoising Diffusion Bridge Models29 Sep
  3. 2023InfoDiffusion: Representation Learning Using Information Maximizing Diffusion Models14 Jun
  4. 2023Improved Techniques for Maximum Likelihood Estimation for Diffusion ODEs6 May
  5. 2023Stable Target Field for Reduced Variance Score Estimation in Diffusion Models1 Feb
  6. 2022Refining Generative Process with Discriminator Guidance in Score-based Diffusion Models28 Nov
  7. 2022FP-Diffusion: Improving Score-based Diffusion Models by Enforcing the Underlying Score Fokker-Planck Equation9 Oct
  8. 2022Classifier-Free Diffusion Guidance26 Jul
  9. 2022EGSDE: Unpaired Image-to-Image Translation via Energy-Guided Stochastic Differential Equations14 Jul
  10. 2022Diffusion models as plug-and-play priors17 Jun
  11. 2022Maximum Likelihood Training for Score-Based Diffusion ODEs by High-Order Denoising Score Matching16 Jun
  12. 2022Elucidating the Design Space of Diffusion-Based Generative Models1 Jun
9 more
  1. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
  2. 2021Tackling the Generative Learning Trilemma with Denoising Diffusion GANs15 Dec
  3. 2021Soft Truncation: A Universal Training Technique of Score-based Diffusion Model for High Precision Score Estimation10 Jun
  4. 2021Diffusion Models Beat GANs on Image Synthesis11 May
  5. 2021Improved Denoising Diffusion Probabilistic Models18 Feb
  6. 2021Maximum Likelihood Training of Score-Based Diffusion Models22 Jan
  7. 2020Score-Based Generative Modeling through Stochastic Differential Equations26 Nov
  8. 2020Denoising Diffusion Probabilistic Models19 Jun
  9. 2020Improved Techniques for Training Score-Based Generative Models16 Jun

Other references · 33

  1. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  2. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
  3. 2014Deep Learning Face Attributes in the Wildcited by 3,679
  4. 2021GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Modelscited by 3,595
  5. 2022DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generationcited by 3,286
  6. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,609
  7. 2016Equality of Opportunity in Supervised Learningcited by 2,000
  8. 2021SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equationscited by 1,903
  9. 2020Training Generative Adversarial Networks with Limited Datacited by 1,184
  10. 2019Learning Imbalanced Datasets with Label-Distribution-Aware Margin Losscited by 987
  11. 2014Certifying and removing disparate impactcited by 715
  12. 2018Learning to Reweight Examples for Robust Deep Learningcited by 670
21 more
  1. 2016f-GAN: Training Generative Neural Samplers using Variational Divergence Minimizationcited by 504
  2. 2017Towards Principled Methods for Training Generative Adversarial Networkscited by 460
  3. 2021How to Train Your Energy-Based Modelscited by 271
  4. 2015The Variational Fair Autoencodercited by 223
  5. 2022Diffusion-GAN: Training GANs with Diffusioncited by 203
  6. 2018FairGAN: Fairness-aware Generative Adversarial Networkscited by 164
  7. 2023Fair Diffusion: Instructing Text-to-Image Generation Models on Fairnesscited by 149
  8. 2017Stabilizing Training of Generative Adversarial Networks through Regularizationcited by 118
  9. 2019Fair Generative Modeling via Weak Supervisioncited by 114
  10. 2020Telescoping Density-Ratio Estimationcited by 108
  11. 201850 Years of Test (Un)fairness: Lessons for Machine Learningcited by 101
  12. 2021Influence-Balanced Loss for Imbalanced Visual Classificationcited by 91
  13. 2015A Deeper Look at Dataset Biascited by 86
  14. 2018Learning Controllable Fair Representationscited by 83
  15. 2022A Systematic Study of Bias Amplificationcited by 65
  16. 2018Fairness Behind a Veil of Ignorance: A Welfare Analysis for Automated Decision Makingcited by 46
  17. 2019AdaFair: Cumulative Fairness Adaptive Boostingcited by 45
  18. 2022Fair Generative Models via Transfer Learningcited by 27
  19. 2016Generative Adversarial Nets from a Density Ratio Estimation Perspectivecited by 24
  20. 2022Learning to Re-weight Examples with Optimal Transport for Imbalanced Classificationcited by 23
  21. 2022Adaptive Multi-stage Density Ratio Estimation for Learning Latent Space Energy-based Modelcited by 18

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 8; 2025: 13; 2026: 4.

25 papers.

Sort:

By area:

  1. 2024SimGen: Simulator-conditioned Driving Scene Generationcited by 34
  2. 2024Constrained Diffusion Models via Dual Trainingcited by 19
  3. 2024Diffusion Rejection Samplingcited by 11
  4. 2025Preference Optimization by Estimating the Ratio of the Data Distributioncited by 9
  5. 2024Diffusion Bridge AutoEncoders for Unsupervised Representation Learningcited by 6
  6. 2025On Fairness of Unified Multimodal Large Language Model for Image Generationcited by 6
  7. 2024PFGuard: A Generative Framework with Privacy and Fairness Safeguardscited by 3
  8. 2025Diffusing DeBias: Synthetic Bias Amplification for Model Debiasingcited by 3
  9. 2025Multi-Group Proportional Representation for Text-to-Image Modelscited by 3
  10. 2025Fully Unsupervised Self-debiasing of Text-to-Image Diffusion Modelscited by 3
  11. 2024ED-SAM: An Efficient Diffusion Sampling Approach to Domain Generalization in Vision-Language Foundation Modelscited by 2
  12. 2024Binary Losses for Density Ratio Estimationcited by 2
13 more
  1. 2025PoGDiff: Product-of-Gaussians Diffusion Models for Imbalanced Text-to-Image Generationcited by 2
  2. 2025SafetyPairs: Isolating Safety Critical Image Features with Counterfactual Image Generationcited by 2
  3. 2025Critical Confabulation: Can LLMs Hallucinate for Social Good?cited by 2
  4. 2026Attention, May I Have Your Decision? Localizing Generative Choices in Diffusion Modelscited by 2
  5. 2026Stay Fair! Ensuring Group Fairness in Diffusion Models Across Guidance Scalescited by 2
  6. 2024Distribution Backtracking Builds A Faster Convergence Trajectory for Diffusion Distillationcited by 1
  7. 2025How Bias Binds: Measuring Hidden Associations for Bias Control in Text-to-Image Compositionscited by 1
  8. 2026Equalized Generative Treatment: Matching f-divergences for Fairness in Generative Modelscited by 1
  9. 2026DASH: Dual-Branch Score Distillation for Guidance-Calibrated Compact Diffusion Modelscited by 1
  10. 2025FairT2I: Mitigating Social Bias in Text-to-Image Generation via Large Language Model-Assisted Detection and Attribute Rebalancingcited by 0
  11. 2025Efficient Weighted Sampling via Score-based Generative Modelscited by 0
  12. 2025A Critical Review of Predominant Bias in Neural Networkscited by 0
  13. 2025A Responsible Face Recognition Approach for Small and Mid-Scale Systems Through Personalized Neural Networkscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2023Unmasking Bias in Diffusion Model Training
  2. 2026Distilling Diffusion Score Discrepancy for Efficient Training Data Attribution
  3. 2025Importance Weighted Score Matching for Diffusion Samplers with Enhanced Mode Coverage
  4. 2026Diffusion models recover accurate mixture weights despite score function insensitivity
  5. 2024InvDiff: Invariant Guidance for Bias Mitigation in Diffusion Models
  6. 2026Generalization Properties of Score-matching Diffusion Models for Intrinsically Low-dimensional Data
  7. 2025Generalization through variance: how noise shapes inductive biases in diffusion models
  8. 2025Deeper Diffusion Models Amplify Bias