Causal Discovery with Metadata-Informed Latent Types

Philippe Brouillard, Alexandre Drouin, Dhanya Sridhar
Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, PMLR 337:785-818, 2026.

Abstract

Causal discovery seeks to recover causal structure from data but the underlying graph is typically identifiable only up to its {Markov} equivalence class. Yet real-world systems often exhibit redundancy, where groups of variables share similar causal roles. We introduce a {Bayesian} causal discovery framework that leverages variable-level metadata to infer latent types and constrain causal interactions across variables. We model causal graphs as type-consistent DAGs and propose t-{DiBS}, a fully differentiable method that jointly learns variable types, graph structure, and metadata representations. Our approach enables principled uncertainty quantification and integrates expressive neural models for metadata. We provide theoretical results showing that, under structured assumptions, metadata combined with typing can improve identifiability beyond classical limits. Empirically, we demonstrate improved performance over standard causal discovery methods on synthetic and pseudo-real datasets, with detailed analysis demonstrating the benefit of joint type and structure learning. These results establish metadata-driven typing as a principled approach to identifiable causal discovery.

Cite this Paper

BibTeX
@InProceedings{pmlr-v337-brouillard26a, title = {Causal Discovery with Metadata-Informed Latent Types}, author = {Brouillard, Philippe and Drouin, Alexandre and Sridhar, Dhanya}, booktitle = {Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence}, pages = {785--818}, year = {2026}, editor = {Perković, Emilija and Malinsky, Daniel}, volume = {337}, series = {Proceedings of Machine Learning Research}, month = {17--21 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v337/main/assets/brouillard26a/brouillard26a.pdf}, url = {https://proceedings.mlr.press/v337/brouillard26a.html}, abstract = {Causal discovery seeks to recover causal structure from data but the underlying graph is typically identifiable only up to its {Markov} equivalence class. Yet real-world systems often exhibit redundancy, where groups of variables share similar causal roles. We introduce a {Bayesian} causal discovery framework that leverages variable-level metadata to infer latent types and constrain causal interactions across variables. We model causal graphs as type-consistent DAGs and propose t-{DiBS}, a fully differentiable method that jointly learns variable types, graph structure, and metadata representations. Our approach enables principled uncertainty quantification and integrates expressive neural models for metadata. We provide theoretical results showing that, under structured assumptions, metadata combined with typing can improve identifiability beyond classical limits. Empirically, we demonstrate improved performance over standard causal discovery methods on synthetic and pseudo-real datasets, with detailed analysis demonstrating the benefit of joint type and structure learning. These results establish metadata-driven typing as a principled approach to identifiable causal discovery.} }
Endnote
%0 Conference Paper %T Causal Discovery with Metadata-Informed Latent Types %A Philippe Brouillard %A Alexandre Drouin %A Dhanya Sridhar %B Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Emilija Perković %E Daniel Malinsky %F pmlr-v337-brouillard26a %I PMLR %P 785--818 %U https://proceedings.mlr.press/v337/brouillard26a.html %V 337 %X Causal discovery seeks to recover causal structure from data but the underlying graph is typically identifiable only up to its {Markov} equivalence class. Yet real-world systems often exhibit redundancy, where groups of variables share similar causal roles. We introduce a {Bayesian} causal discovery framework that leverages variable-level metadata to infer latent types and constrain causal interactions across variables. We model causal graphs as type-consistent DAGs and propose t-{DiBS}, a fully differentiable method that jointly learns variable types, graph structure, and metadata representations. Our approach enables principled uncertainty quantification and integrates expressive neural models for metadata. We provide theoretical results showing that, under structured assumptions, metadata combined with typing can improve identifiability beyond classical limits. Empirically, we demonstrate improved performance over standard causal discovery methods on synthetic and pseudo-real datasets, with detailed analysis demonstrating the benefit of joint type and structure learning. These results establish metadata-driven typing as a principled approach to identifiable causal discovery.
APA
Brouillard, P., Drouin, A. & Sridhar, D.. (2026). Causal Discovery with Metadata-Informed Latent Types. Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 337:785-818 Available from https://proceedings.mlr.press/v337/brouillard26a.html.

Related Material