Unified Causal Discovery and Missing Data Imputation

Osman Mian, Jens Kleesiek, Michael Kamp
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:2206-2214, 2026.

Abstract

Causal discovery and data imputation are often treated separately, yet both face challenges when data are missing. Existing causal discovery methods discard incomplete samples, losing valuable information, while standard imputation relies on spurious correlations that obscure the causal signal. We propose LOGIC, a framework that performs causal discovery and causally consistent imputation jointly. In contrast to prior work that assumes all source variables are observed, we derive a verifiable criterion for this assumption under MCAR and MAR missingness, grounded in the Algorithmic Markov Condition. LOGIC then proceeds layer by layer: identifying sources, recovering downstream relations, and imputing missing values, while explicitly declaring unknowns when imputation is unsupported. This design preserves causal reasoning even in challenging missingness regimes. Experiments on synthetic and real-world data show that LOGIC outperforms state-of-the-art baselines in both structure recovery and imputation accuracy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v300-mian26a, title = { Unified Causal Discovery and Missing Data Imputation }, author = {Mian, Osman and Kleesiek, Jens and Kamp, Michael}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {2206--2214}, year = {2026}, editor = {Khan, Emtiyaz and Li, Yingzhen and Solin, Arno and Ramdas, Aaditya}, volume = {300}, series = {Proceedings of Machine Learning Research}, month = {02--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v300/main/assets/mian26a/mian26a.pdf}, url = {https://proceedings.mlr.press/v300/mian26a.html}, abstract = { Causal discovery and data imputation are often treated separately, yet both face challenges when data are missing. Existing causal discovery methods discard incomplete samples, losing valuable information, while standard imputation relies on spurious correlations that obscure the causal signal. We propose LOGIC, a framework that performs causal discovery and causally consistent imputation jointly. In contrast to prior work that assumes all source variables are observed, we derive a verifiable criterion for this assumption under MCAR and MAR missingness, grounded in the Algorithmic Markov Condition. LOGIC then proceeds layer by layer: identifying sources, recovering downstream relations, and imputing missing values, while explicitly declaring unknowns when imputation is unsupported. This design preserves causal reasoning even in challenging missingness regimes. Experiments on synthetic and real-world data show that LOGIC outperforms state-of-the-art baselines in both structure recovery and imputation accuracy. } }
Endnote
%0 Conference Paper %T Unified Causal Discovery and Missing Data Imputation %A Osman Mian %A Jens Kleesiek %A Michael Kamp %B Proceedings of The 29th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2026 %E Emtiyaz Khan %E Yingzhen Li %E Arno Solin %E Aaditya Ramdas %F pmlr-v300-mian26a %I PMLR %P 2206--2214 %U https://proceedings.mlr.press/v300/mian26a.html %V 300 %X Causal discovery and data imputation are often treated separately, yet both face challenges when data are missing. Existing causal discovery methods discard incomplete samples, losing valuable information, while standard imputation relies on spurious correlations that obscure the causal signal. We propose LOGIC, a framework that performs causal discovery and causally consistent imputation jointly. In contrast to prior work that assumes all source variables are observed, we derive a verifiable criterion for this assumption under MCAR and MAR missingness, grounded in the Algorithmic Markov Condition. LOGIC then proceeds layer by layer: identifying sources, recovering downstream relations, and imputing missing values, while explicitly declaring unknowns when imputation is unsupported. This design preserves causal reasoning even in challenging missingness regimes. Experiments on synthetic and real-world data show that LOGIC outperforms state-of-the-art baselines in both structure recovery and imputation accuracy.
APA
Mian, O., Kleesiek, J. & Kamp, M.. (2026). Unified Causal Discovery and Missing Data Imputation . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:2206-2214 Available from https://proceedings.mlr.press/v300/mian26a.html.

Related Material