Incorporating Expert Knowledge into Bayesian Causal Discovery of Mixtures of Directed Acyclic Graphs

Zachris Björkman, Jorge Loria, Sophie Wharrie, Samuel Kaski
Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, PMLR 300:964-972, 2026.

Abstract

Bayesian causal discovery benefits from prior information elicited from domain experts, and in heterogeneous domains any prior knowledge would be badly needed. However, so far prior elicitation approaches have assumed a single causal graph and hence are not suited to heterogeneous domains. We propose a causal elicitation strategy for heterogeneous settings, based on Bayesian experimental design (BED) principles, and a \emph{variational mixture structure learning} (VaMSL) method—extending the earlier \emph{differentiable Bayesian structure learning} (DiBS) method—to iteratively infer mixtures of causal Bayesian networks (CBNs). We construct an informative graph prior incorporating elicited expert feedback in the inference of mixtures of CBNs. Our proposed method successfully produces a set of alternative causal models (mixture components or clusters), and achieves an improved structure learning performance on heterogeneous synthetic data when informed by a simulated expert. Finally, we demonstrate that our approach is capable of capturing complex distributions in a breast cancer database.

Cite this Paper

BibTeX
@InProceedings{pmlr-v300-bjorkman26a, title = { Incorporating Expert Knowledge into Bayesian Causal Discovery of Mixtures of Directed Acyclic Graphs }, author = {Bj{\"o}rkman, Zachris and Loria, Jorge and Wharrie, Sophie and Kaski, Samuel}, booktitle = {Proceedings of The 29th International Conference on Artificial Intelligence and Statistics}, pages = {964--972}, 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/bjorkman26a/bjorkman26a.pdf}, url = {https://proceedings.mlr.press/v300/bjorkman26a.html}, abstract = { Bayesian causal discovery benefits from prior information elicited from domain experts, and in heterogeneous domains any prior knowledge would be badly needed. However, so far prior elicitation approaches have assumed a single causal graph and hence are not suited to heterogeneous domains. We propose a causal elicitation strategy for heterogeneous settings, based on Bayesian experimental design (BED) principles, and a \emph{variational mixture structure learning} (VaMSL) method—extending the earlier \emph{differentiable Bayesian structure learning} (DiBS) method—to iteratively infer mixtures of causal Bayesian networks (CBNs). We construct an informative graph prior incorporating elicited expert feedback in the inference of mixtures of CBNs. Our proposed method successfully produces a set of alternative causal models (mixture components or clusters), and achieves an improved structure learning performance on heterogeneous synthetic data when informed by a simulated expert. Finally, we demonstrate that our approach is capable of capturing complex distributions in a breast cancer database. } }
Endnote
%0 Conference Paper %T Incorporating Expert Knowledge into Bayesian Causal Discovery of Mixtures of Directed Acyclic Graphs %A Zachris Björkman %A Jorge Loria %A Sophie Wharrie %A Samuel Kaski %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-bjorkman26a %I PMLR %P 964--972 %U https://proceedings.mlr.press/v300/bjorkman26a.html %V 300 %X Bayesian causal discovery benefits from prior information elicited from domain experts, and in heterogeneous domains any prior knowledge would be badly needed. However, so far prior elicitation approaches have assumed a single causal graph and hence are not suited to heterogeneous domains. We propose a causal elicitation strategy for heterogeneous settings, based on Bayesian experimental design (BED) principles, and a \emph{variational mixture structure learning} (VaMSL) method—extending the earlier \emph{differentiable Bayesian structure learning} (DiBS) method—to iteratively infer mixtures of causal Bayesian networks (CBNs). We construct an informative graph prior incorporating elicited expert feedback in the inference of mixtures of CBNs. Our proposed method successfully produces a set of alternative causal models (mixture components or clusters), and achieves an improved structure learning performance on heterogeneous synthetic data when informed by a simulated expert. Finally, we demonstrate that our approach is capable of capturing complex distributions in a breast cancer database.
APA
Björkman, Z., Loria, J., Wharrie, S. & Kaski, S.. (2026). Incorporating Expert Knowledge into Bayesian Causal Discovery of Mixtures of Directed Acyclic Graphs . Proceedings of The 29th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 300:964-972 Available from https://proceedings.mlr.press/v300/bjorkman26a.html.

Related Material