mlsanityData to 14 Oct 2026

MissDiff: Training Diffusion Models on Tabular Data with Missing Values

arXiv 2307.00467 ↗First posted 2 Jul 2023cs.LG stat.MLcited by 39

On the map: Diffusion models → Flow matching, time series and science

Abstract (arXiv)

The diffusion model has shown remarkable performance in modeling data distributions and synthesizing data. However, the vanilla diffusion model requires complete or fully observed data for training. Incomplete data is a common issue in various real-world applications, including healthcare and finance, particularly when dealing with tabular datasets. This work presents a unified and principled diffusion-based framework for learning from data with missing values under various missing mechanisms. We first observe that the widely adopted "impute-then-generate" pipeline may lead to a biased learning objective. Then we propose to mask the regression loss of Denoising Score Matching in the training phase. We prove the proposed method is consistent in learning the score of data distributions, and the proposed training objective serves as an upper bound for the negative likelihood in certain cases. The proposed framework is evaluated on multiple tabular datasets using realistic and efficacious metrics and is demonstrated to outperform state-of-the-art diffusion model on tabular data with "impute-then-generate" pipeline by a large margin.

Key facts. Cited by 39 papers in the map (data through 14 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 37 of its 73 references that are in the map. First posted on arXiv 2 Jul 2023.

Builds on

37 of 73 references.

Closest ancestors · 4

Same direction, newest first

  1. 2022Diffusion models for missing value imputation in tabular data31 Oct
  2. 2022STaSy: Score-based Tabular data Synthesis8 Oct
  3. 2022TabDDPM: Modelling Tabular Data with Diffusion Models30 Sep
  4. 2021CSDI: Conditional Score-based Diffusion Models for Probabilistic Time Series Imputation7 Jul

Other references · 33

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  3. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
  4. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,255
  5. 2021Diffusion Probabilistic Models for 3D Point Cloud Generationcited by 735
  6. 2019Modeling Tabular data using Conditional GANcited by 699
  7. 2021Maximum Likelihood Training of Score-Based Diffusion Modelscited by 672
  8. 2019Sliced Score Matching: A Scalable Approach to Density and Score Estimationcited by 384
  9. 2018GAIN: Missing Data Imputation using Generative Adversarial Netscited by 352
  10. 2023Synthetic Data from Diffusion Models Improves ImageNet Classificationcited by 327
  11. 2021Improving Robustness using Generated Datacited by 234
  12. 2020How to train your neural ODE: the world of Jacobian and kinetic regularizationcited by 228
21 more
  1. 2017VEEGAN: Reducing Mode Collapse in GANs using Implicit Variational Learningcited by 200
  2. 2018Data Synthesis based on Generative Adversarial Networkscited by 185
  3. 2022Score-based Continuous-time Discrete Diffusion Modelscited by 169
  4. 2018Handling Incomplete Heterogeneous Data using VAEscited by 149
  5. 2022Concrete Score Matching: Generalized Score Matching for Discrete Datacited by 148
  6. 2021Robust Learning Meets Generative Models: Can Proxy Distributions Improve Adversarial Robustness?cited by 113
  7. 2023Ambient Diffusion: Learning Clean Distributions from Corrupted Datacited by 111
  8. 2020Missing Data Imputation using Optimal Transportcited by 100
  9. 2022HyperImpute: Generalized Iterative Imputation with Automatic Model Selectioncited by 83
  10. 2019MisGAN: Learning from Incomplete Data with Generative Adversarial Networkscited by 75
  11. 2020VAEM: a Deep Generative Model for Heterogeneous Mixed Type Datacited by 66
  12. 2020not-MIWAE: Deep Generative Modelling with Missing not at Random Datacited by 62
  13. 2023Improved Techniques for Maximum Likelihood Estimation for Diffusion ODEscited by 54
  14. 2020Learning from Irregularly-Sampled Time Series: A Missing Data Perspectivecited by 40
  15. 2020MCFlow: Monte Carlo Flow Models for Data Imputationcited by 37
  16. 2023Diffusion Models and Semi-Supervised Learners Benefit Mutually with Few Labelscited by 37
  17. 2021Benchmarking emergency department triage prediction models with machine learning and large public electronic health recordscited by 31
  18. 2021OCT-GAN: Neural ODE-based Conditional Tabular GANscited by 27
  19. 2016Personalized Risk Scoring for Critical Care Prognosis using Mixtures of Gaussian Processescited by 19
  20. 2020PC-GAIN: Pseudo-label Conditional Generative Adversarial Imputation Networks for Incomplete Datacited by 15
  21. 2017General Latent Feature Models for Heterogeneous Datasetscited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 2; 2024: 9; 2025: 17; 2026: 11.

39 papers.

Sort:

By area:

  1. 2024An Overview of Diffusion Models: Applications, Guided Generation, Statistical Rates and Optimizationcited by 73
  2. 2024A Survey on Diffusion Models for Time Series and Spatio-Temporal Datacited by 67
  3. 2023Generating and Imputing Tabular Data via Diffusion and Flow-based Gradient-Boosted Treescited by 62
  4. 2023FinDiff: Diffusion Models for Financial Tabular Data Generationcited by 41
  5. 2025A Comprehensive Survey of Synthetic Tabular Data Generationcited by 28
  6. 2024A Survey on Statistical Theory of Deep Learning: Approximation, Training Dynamics, and Generative Modelscited by 20
  7. 2024DiffPuter: Empowering Diffusion Models for Missing Data Imputationcited by 19
  8. 2024Self-Supervision Improves Diffusion Models for Tabular Data Imputationcited by 16
  9. 2025Diffusion and Flow Matching Models for Tabular Data: A Surveycited by 8
  10. 2025Score Matching With Missing Datacited by 8
  11. 2025Anomaly Detection and Generation with Diffusion Models: A Surveycited by 8
  12. 2024DiffImpute: Tabular Data Imputation With Denoising Diffusion Probabilistic Modelcited by 6
27 more
  1. 2025Missing Data Imputation by Reducing Mutual Information with Rectified Flowscited by 6
  2. 2024Rethinking the Diffusion Models for Numerical Tabular Data Imputation from the Perspective of Wasserstein Gradient Flowcited by 3
  3. 2025Incomplete Data, Complete Dynamics: A Diffusion Approachcited by 3
  4. 2025A diffusion-based generative model for financial time series via geometric Brownian motioncited by 2
  5. 2025Impute-MACFM: Imputation based on Mask-Aware Flow Matchingcited by 2
  6. 2024Risk-Sensitive Diffusion: Robustly Optimizing Diffusion Models with Noisy Samplescited by 1
  7. 2025DeepIFSAC: Deep Imputation of Missing Values Using Feature and Sample Attention within Contrastive Frameworkcited by 1
  8. 2025Imputation-free Learning of Tabular Data with Missing Values using Incremental Feature Partitions in Transformercited by 1
  9. 2025How Far Are We from Generating Missing Modalities with Foundation Models?cited by 1
  10. 2026CoreFlow: Low-Rank Matrix Generative Modelscited by 1
  11. 2026AugMask: Training Diffusion Models on Incomplete Tabular Data via Stochastic Augmentation and Maskingcited by 1
  12. 2024Watermarking Generative Categorical Datacited by 0
  13. 2025Kriging and Gaussian Process Interpolation for Georeferenced Data Augmentationcited by 0
  14. 2025Interpolation pour l'augmentation de donnees : Application `a la gestion des adventices de la canne a sucre a la Reunioncited by 0
  15. 2025Diffusion Transformers for Imputation: Statistical Efficiency and Uncertainty Quantificationcited by 0
  16. 2025Learning Robust Diffusion Models from Imprecise Supervisioncited by 0
  17. 2025An Interdisciplinary and Cross-Task Review on Missing Data Imputationcited by 0
  18. 2025Boosting Predictive Performance on Tabular Data through Data Augmentation with Latent-Space Flow-Based Diffusioncited by 0
  19. 2026Harpoon: Generalised Manifold Guidance for Conditional Tabular Diffusioncited by 0
  20. 2026Generative Modeling under Non-Monotone MAR Missingness via Approximate Wasserstein Gradient Flowscited by 0
  21. 2026Observation-Aligned Mask Priors for Learning Physical Fields from Authentic Occlusionscited by 0
  22. 2026TabKDE: Simple and Scalable Tabular Data Generation with Kernel Density Estimatescited by 0
  23. 2026Latent Diffusion for Missing Datacited by 0
  24. 2026Learning What Not to Impute: An Uncertainty-Aware Diffusion Framework for Meaningful Missingnesscited by 0
  25. 2026TabSODA: Tabular Diffusion based Imputation with Skip Pattern Detection and Ordinal Awarenesscited by 0
  26. 2026Impute-EM: Native Mixed-State Diffusion Models for Heterogeneous Data Imputationcited by 0
  27. 2026Learning Field Reconstruction from Incomplete Data by Globally Correcting Local Estimatescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025MissDDIM: Deterministic and Efficient Conditional Diffusion for Tabular Data Imputation
  2. 2024Diffusion Models for Tabular Data Imputation and Synthetic Data Generation
  3. 2024TabDiff: a Mixed-type Diffusion Model for Tabular Data Generation
  4. 2025MissHDD: Hybrid Deterministic Diffusion for Hetrogeneous Incomplete Data Imputation
  5. 2026CDMD: A Cross-Dataset Mixed-Type Diffusion Model for Tabular Data
  6. 2023Consistent Diffusion Models: Mitigating Sampling Drift by Learning to be Consistent
  7. 2026TabMosaic: Unified In-Context Diffusion for Tabular Generative Tasks
  8. 2023Continuous Diffusion for Mixed-Type Tabular Data