mlsanityData to 7 Oct 2026

Bias Correction of Learned Generative Models using Likelihood-Free Importance Weighting

arXiv 1906.09531 ↗First posted 23 Jun 2019stat.ML cs.LG cs.NEcited by 70

Not on the map. Why?

Abstract (arXiv)

A learned generative model often produces biased statistics relative to the underlying data distribution. A standard technique to correct this bias is importance sampling, where samples from the model are weighted by the likelihood ratio under model and true distributions. When the likelihood ratio is unknown, it can be estimated by training a probabilistic classifier to distinguish samples from the two distributions. We employ this likelihood-free importance weighting method to correct for the bias in generative models. We find that this technique consistently improves standard goodness-of-fit metrics for evaluating the sample quality of state-of-the-art deep generative models, suggesting reduced bias. Finally, we demonstrate its utility on representative applications in a) data augmentation for classification using generative adversarial networks, and b) model-based policy evaluation using off-policy data.

Key facts. Cited by 70 papers in the map (data through 7 Oct 2026). First posted on arXiv 23 Jun 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 17; 2022: 19; 2023: 18; 2024: 6; 2025: 8; 2026: 2.

70 papers.

Sort:

By area:

  1. 2023Steering Language Models With Activation Engineeringcited by 824
  2. 2021Pretrained Transformers as Universal Computation Enginescited by 162
  3. 2021EditGAN: High-Precision Semantic Image Editingcited by 156
  4. 2023Debiasing Vision-Language Models via Biased Promptscited by 147
  5. 2022StyleGAN-Human: A Data-Centric Odyssey of Human Generationcited by 143
  6. 2022A Survey of Controllable Text Generation using Transformer-based Pre-trained Language Modelscited by 128
  7. 2022A Prompt Array Keeps the Bias Away: Debiasing Vision-Language Models with Adversarial Learningcited by 104
  8. 2022How to Backdoor Diffusion Models?cited by 102
  9. 2023Art and the science of generative AI: A deeper divecited by 93
  10. 2022RankGen: Improving Text Generation with Large Ranking Modelscited by 72
  11. 2023VillanDiffusion: A Unified Backdoor Attack Framework for Diffusion Modelscited by 68
  12. 2023Group Preference Optimization: Few-Shot Alignment of Large Language Modelscited by 61
38 more
  1. 2021Mandoline: Model Evaluation under Distribution Shiftcited by 55
  2. 2024Fairness Feedback Loops: Training on Synthetic Data Amplifies Biascited by 55
  3. 2021Don't Generate Me: Training Differentially Private Generative Models with Sinkhorn Divergencecited by 54
  4. 2022Leveling Down in Computer Vision: Pareto Inefficiencies in Fair Deep Classifierscited by 53
  5. 2023Stable Target Field for Reduced Variance Score Estimation in Diffusion Modelscited by 44
  6. 2023Mass-Producing Failures of Multimodal Systems with Language Modelscited by 36
  7. 2024Self-Consuming Generative Models with Curated Data Provably Optimize Human Preferencescited by 35
  8. 2021Featurized Density Ratio Estimationcited by 33
  9. 2021Measuring Fairness in Generative Modelscited by 31
  10. 2022Counterfactually Measuring and Eliminating Social Bias in Vision-Language Pre-training Modelscited by 28
  11. 2023Tailoring Language Generation Models under Total Variation Distancecited by 27
  12. 2023Survey of Social Bias in Vision-Language Modelscited by 26
  13. 2022Polarity Sampling: Quality and Diversity Control of Pre-Trained Generative Networks via Singular Valuescited by 21
  14. 2022FairStyle: Debiasing StyleGAN2 with Style Channel Manipulationscited by 20
  15. 2023EHRDiff: Exploring Realistic EHR Synthesis with Diffusion Modelscited by 18
  16. 2023On the Trustworthiness Landscape of State-of-the-art Generative Models: A Survey and Outlookcited by 15
  17. 2024How to Leverage Demonstration Data in Alignment for Large Language Model? A Self-Imitation Learning Perspectivecited by 14
  18. 2021Self-Diagnosing GAN: Diagnosing Underrepresented Samples in Generative Adversarial Networkscited by 13
  19. 2021Unaligned Image-to-Image Translation by Learning to Reweightcited by 13
  20. 2022Controlling Directions Orthogonal to a Classifiercited by 13
  21. 2021AI Fairness via Domain Adaptationcited by 8
  22. 2023Synthetic Data in Healthcarecited by 8
  23. 2023Time-series Generation by Contrastive Imitationcited by 8
  24. 2023Advancements in Scientific Controllable Text Generation Methodscited by 7
  25. 2021Latent reweighting, an almost free improvement for GANscited by 6
  26. 2022Self-Conditioned Generative Adversarial Networks for Image Editingcited by 6
  27. 2023Towards Mode Balancing of Generative Models via Diversity Weightscited by 6
  28. 2021A Unified Framework for Multi-distribution Density Ratio Estimationcited by 5
  29. 2022Long-term Control for Dialogue Generation: Methods and Evaluationcited by 5
  30. 2022Breaking the Spurious Causality of Conditional Generation via Fairness Intervention with Corrective Samplingcited by 5
  31. 2025Steering Risk Preferences in Large Language Models by Aligning Behavioral and Neural Representationscited by 5
  32. 2021Mitigating Statistical Bias within Differentially Private Synthetic Datacited by 4
  33. 2023Gaussian Harmony: Attaining Fairness in Diffusion-based Face Generation Modelscited by 4
  34. 2022Imitating, Fast and Slow: Robust learning from demonstrations via decision-time planningcited by 3
  35. 2021Discriminator Augmented Model-Based Reinforcement Learningcited by 2
  36. 2021Fair Visual Recognition in Limited Data Regime using Self-Supervision and Self-Distillationcited by 2
  37. 2022Gradient Based Activations for Accurate Bias-Free Learningcited by 2
  38. 2023Is novelty predictable?cited by 2

50 most cited of 70 All 70 (JSON)