Research map / Probabilistic and scientific modeling / Causal discovery
MIRACLE: Causally-Aware Imputation via Learning Missing Data Mechanisms
Trent Kyono, Yao Zhang, Alexis Bellot, Mihaela van der Schaar
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Refine MissForest imputations with MIRACLE | MissForest · Synthetic X5; missingness caused by X2 (MAR) | RMSE: 1.31 ± .06 -> 0.49 ± .06 | Improved |
| Refine MissForest imputations with MIRACLE | MissForest · Synthetic X5; missingness caused by X3 (MAR) | RMSE: 1.45 ± .09 -> 0.50 ± .06 | Improved |
| Refine MissForest imputations with MIRACLE | MissForest · Synthetic X5; missingness caused by X5 (MNAR) | RMSE: 1.76 ± .06 -> 1.59 ± .07 | Improved |
| Refine MissForest imputations with MIRACLE | MissForest · Synthetic X5; missingness caused by X9 (MCAR) | RMSE: 0.17 ± .07 -> 0.14 ± .05 | Improved |
| Refine MissForest imputations with MIRACLE | MissForest · Synthetic X5; missingness caused by X4 (MAR) | RMSE: 1.23 ± .05 -> 1.04 ± .05 | Improved |
| Remove reconstruction loss L1 (ablation) | MIRACLE · Abalone, MAR | RMSE: 0.222 ± 0.062 -> 0.321 ± 0.108 | Hurt |
| Remove causal regularizer R1 (ablation) | MIRACLE · Abalone, MAR | RMSE: 0.222 ± 0.062 -> 0.521 ± 0.199 | Hurt |
| Remove moment regularizer R2 (ablation) | MIRACLE · Abalone, MAR | RMSE: 0.222 ± 0.062 -> 0.312 ± 0.082 | Hurt |
| Use all three loss terms L1, R1, and R2 rather than any reported two-term variant (ablation) | MIRACLE · Real datasets, MAR; MissForest baseline | RMSE | Improved |
| Increase dataset size for learning the missingness graph | MIRACLE · Synthetic data | DAG weight quality: 0.135 ± 0.04; 0.242 ±0.03; 0.252 ±0.03; 0.910 ±0.01; 0.916 ±0.01 | Improved |
Closest papers by meaning
- Learning What Not to Impute: An Uncertainty-Aware Diffusion Framework for Meaningful Missingness (NeurIPS 2026 · similarity 0.88)
- Distribution Shift in Missing Data Imputation: A Risk-Based Perspective and Importance-Weighted Correction under MAR (NeurIPS 2026 · similarity 0.87)
- Identification of Causal Structure in the Presence of Missing Data with Additive Noise Model (AAAI 2024 · similarity 0.86)
- Random features models: a way to study the success of naive imputation (ICML 2024 · similarity 0.86)
- Prediction models that learn to avoid missing values (ICML 2025 · similarity 0.85)
- Missing Data Imputation by Reducing Mutual Information with Rectified Flows (NeurIPS 2025 · 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.
