mlsanityData to 14 Oct 2026

DoubleGen: Debiased Generative Modeling of Counterfactuals

arXiv 2509.16842 ↗First posted 20 Sep 2025stat.ML cs.LG stat.MEcited by 9

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Generative models for counterfactual outcomes face two key sources of bias. Confounding bias arises when approaches fail to account for systematic differences between those who receive the intervention and those who do not. Misspecification bias arises when methods attempt to address confounding through estimation of an auxiliary model, but specify it incorrectly. We introduce DoubleGen, a doubly robust framework that modifies generative modeling training objectives to mitigate these biases. The new objectives rely on two auxiliaries -- a propensity and outcome model -- and successfully address confounding bias even if only one of them is correct. We provide finite-sample guarantees for this robustness property. We further establish conditions under which DoubleGen achieves oracle optimality -- matching the convergence rates standard approaches would enjoy if interventional data were available -- and minimax rate optimality. We illustrate DoubleGen with three examples: diffusion models, flow matching, and autoregressive language models.

Key facts. Cited by 9 papers in the map (data through 14 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 60 of its 104 references that are in the map. First posted on arXiv 20 Sep 2025.

Builds on

60 of 104 references.

Closest ancestors · 11

Same direction, newest first

  1. 2024Statistical guarantees for denoising reflected diffusion models3 Nov
  2. 2023Error Bounds for Flow Matching Methods26 May
  3. 2023Diffusion Models are Minimax Optimal Distribution Estimators3 Mar
  4. 2023Score Approximation, Estimation and Distribution Recovery of Diffusion Models on Low-Dimensional Data14 Feb
  5. 2023Modeling Causal Mechanisms with Diffusion Models for Interventional and Counterfactual Queries2 Feb
  6. 2022Convergence of denoising diffusion models under the manifold hypothesis10 Aug
  7. 2022Classifier-Free Diffusion Guidance26 Jul
  8. 2022Convergence for score-based generative modeling with polynomial complexity13 Jun
  9. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
  10. 2021Conditional Image Generation with Score-Based Diffusion Models26 Nov
  11. 2020Score-Based Generative Modeling through Stochastic Differential Equations26 Nov

Other references · 49

  1. 2017Attention Is All You Needcited by 60,361
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  3. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  4. 2024The Llama 3 Herd of Modelscited by 14,715
  5. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  6. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
  7. 2022Flow Matching for Generative Modelingcited by 5,702
  8. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,475
  9. 2021Zero-Shot Text-to-Image Generationcited by 4,288
  10. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,700
  11. 2014Deep Learning Face Attributes in the Wildcited by 3,693
  12. 2019The Curious Case of Neural Text Degenerationcited by 2,916
37 more
  1. 2019Optuna: A Next-generation Hyperparameter Optimization Frameworkcited by 2,501
  2. 2019HuggingFace's Transformers: State-of-the-art Natural Language Processingcited by 2,218
  3. 2021Palette: Image-to-Image Diffusion Modelscited by 1,300
  4. 2018Demystifying MMD GANscited by 1,293
  5. 2018Hierarchical Neural Story Generationcited by 1,193
  6. 2019Improved Precision and Recall Metric for Assessing Generative Modelscited by 944
  7. 2019MediaPipe: A Framework for Building Perception Pipelinescited by 583
  8. 2013Generating Sequences With Recurrent Neural Networkscited by 554
  9. 2021MAUVE: Measuring the Gap Between Neural Text and Human Text using Divergence Frontierscited by 470
  10. 2017Quasi-Oracle Estimation of Heterogeneous Treatment Effectscited by 332
  11. 2024Bridging Language and Items for Retrieval and Recommendation: Benchmarking LLMs as Semantic Encoderscited by 224
  12. 2020Deep Structural Causal Models for Tractable Counterfactual Inferencecited by 211
  13. 2018Adaptivity of deep ReLU network for learning in Besov and mixed smooth Besov spaces: optimal rate and curse of dimensionalitycited by 162
  14. 2017CausalGAN: Learning Causal Implicit Generative Models with Adversarial Trainingcited by 152
  15. 2019Orthogonal Statistical Learningcited by 151
  16. 2020Algorithmic recourse under imperfect causal knowledge: a probabilistic approachcited by 140
  17. 2020Causal Autoregressive Flowscited by 99
  18. 2018Language GANs Falling Shortcited by 97
  19. 2022Diffusion Causal Models for Counterfactual Estimationcited by 87
  20. 2023Non-Vacuous Generalization Bounds for Large Language Modelscited by 53
  21. 2023Causal normalizing flows: from theory to practicecited by 51
  22. 2021Semiparametric counterfactual density estimationcited by 42
  23. 2020A Systematic Characterization of Sampling Algorithms for Open-ended Language Generationcited by 28
  24. 2022Evaluate & Evaluation on the Hub: Better Best Practices for Data and Model Measurementscited by 22
  25. 2023From Identifiable Causal Representations to Controllable Counterfactual Generation: A Survey on Causal Generative Modelingcited by 22
  26. 2021VACA: Design of Variational Graph Autoencoders for Interventional and Counterfactual Queriescited by 19
  27. 2024Combining T-learning and DR-learning: a framework for oracle-efficient estimation of causal contrastscited by 18
  28. 2023Counterfactual Generative Models for Time-Varying Treatmentscited by 17
  29. 2017Causal Inference: A Missing Data Perspectivecited by 15
  30. 2023One-Step Estimation of Differentiable Hilbert-Valued Parameterscited by 11
  31. 2017Sequential Double Robustness in Right-Censored Longitudinal Modelscited by 10
  32. 2023An Efficient Doubly-Robust Test for the Kernel Treatment Effectcited by 10
  33. 2022Fast convergence rates for dose-response estimationcited by 8
  34. 2017On the multiply robust estimation of the mean of the g-functionalcited by 6
  35. 2020A Note on High-Probability versus In-Expectation Guarantees of Generalization Bounds in Machine Learningcited by 3
  36. 2018MMLSpark: Unifying Machine Learning Ecosystems at Massive Scalescited by 2
  37. 2025Stochastic Gradients under Nuisancescited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 8.

9 papers.

  1. 2026Harnessing Synthetic Data from Generative AI for Statistical Inferencecited by 3
  2. 2026Debiased Counterfactual Generation via Flow Matching from Observationscited by 1
  3. 2025Learning density ratios in causal inference using Bregman-Riesz regressioncited by 0
  4. 2026Extended Wasserstein-GAN Approach to Causal Distribution Learning: Density-Free Estimation and Minimax Optimalitycited by 0
  5. 2026Geometry Adaptive Counterfactual Distribution Learning with Diffusion-Guided Smoothingcited by 0
  6. 2026Are Good Generators Good Decision-Makers? Policy Learning for General Interventions via Retargeted Counterfactual Generationcited by 0
  7. 2026Counterfactual Generation via Flow Matching: Coupling-Sensitive End-to-End Ratescited by 0
  8. 2026OrthoGen: A Generative Orthogonal Learner for Time-Varying Treatmentscited by 0
  9. 2026Amortized Counterfactual Prediction with Graph Neural Networkscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026The Fairness Collapse Phenomenon: Bias Amplification in Language Models Trained on Synthetic Data
  2. 2024Bias and Volatility: A Statistical Framework for Evaluating Large Language Model's Stereotypes and the Associated Generation Inconsistency
  3. 2025Deeper Diffusion Models Amplify Bias
  4. 2026CogBias: Measuring and Mitigating Cognitive Bias in Large Language Models
  5. 2025How Quantization Shapes Bias in Large Language Models
  6. 2025Activation Steering for Bias Mitigation: An Interpretable Approach to Safer LLMs
  7. 2024Ablation Based Counterfactuals
  8. 2026QuantiBias: Benchmarking Quantization-Induced Bias in LLMs