Research map / Probabilistic and scientific modeling / Variational and generative sampling
not-MIWAE: Deep Generative Modelling with Missing not at Random Data
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Model the self-masking process with known direction rather than use standard PPCA | PPCA on UCI datasets · Banknote, Concrete, Red, White, Yeast, Breast | imputation RMSE: 1.39 ± 0.00 -> 0.57 ± 0.00; 1.61 ± 0.00 -> 1.31 ± 0.00; 1.61 ± 0.00 -> 1.13 ± 0.00; 1.57 ± 0.00 -> 0.99 ± 0.00; 1.67 ± 0 | Helped |
| Use a self-masking model with known direction rather than an agnostic missing model (ablation) | not-MIWAE - PPCA on UCI datasets · Banknote, Concrete, Red, White, Yeast, Breast | imputation RMSE: 1.25 ± 0.15 -> 0.57 ± 0.00; 1.47 ± 0.01 -> 1.31 ± 0.00; 1.32 ± 0.00 -> 1.13 ± 0.00; 1.27 ± 0.01 -> 0.99 ± 0.00; 1.20 ± 0 | Helped |
| Use a self-masking model with known direction rather than an agnostic missing model (ablation) | not-MIWAE on UCI datasets · Banknote, Concrete, Red, White, Yeast, Breast | imputation RMSE: 0.80 ± 0.08 -> 0.74 ± 0.05; 2.63 ± 0.12 -> 1.12 ± 0.04; 1.30 ± 0.01 -> 1.07 ± 0.00; 1.37 ± 0.00 -> 1.04 ± 0.00; 1.43 ± 0 | Mixed |
| Use a flexible deep data model with an agnostic missing model instead of a PPCA-like data model | not-MIWAE on UCI datasets · Concrete and Yeast | imputation RMSE: 1.47 ± 0.01 -> 2.63 ± 0.12; 1.20 ± 0.05 -> 1.43 ± 0.02 | Hurt |
| Jointly model the missing process rather than ignore it | not-MIWAE Gaussian versus MIWAE Gaussian · Yahoo! R3 MCAR test-set | imputation MSE: 2.055 ± 0.001 -> 0.939 ± 0.007 | Helped |
| Use Gaussian rather than categorical observation model (ablation) | not-MIWAE on Yahoo! R3 · Yahoo! R3 MCAR test-set | imputation MSE: 1.293 ± 0.006 -> 0.939 ± 0.007 | Helped |
| Model MNAR selection jointly without using propensity estimates | not-MIWAE Gaussian versus propensity-based baselines · Yahoo! R3 MCAR test-set | imputation MSE: 0.966 -> 0.939 ± 0.007; 0.957 -> 0.939 ± 0.007 | Helped |
Closest papers by meaning
- Latent variable model for high-dimensional point process with structured missingness (ICML 2024 · similarity 0.87)
- Generative Modeling under Non-Monotone MAR Missingness via Approximate Wasserstein Gradient Flows (NeurIPS 2026 · similarity 0.87)
- DiffPuter: Empowering Diffusion Models for Missing Data Imputation (ICLR 2025 · similarity 0.86)
- Prediction models that learn to avoid missing values (ICML 2025 · similarity 0.85)
- A Likelihood Based Approach to Distribution Regression Using Conditional Deep Generative Models (ICML 2025 · similarity 0.85)
- Learning What Not to Impute: An Uncertainty-Aware Diffusion Framework for Meaningful Missingness (NeurIPS 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.