Causal Preference Elicitation
This work addresses the challenge of efficient causal discovery for researchers in fields like biology and medicine, offering an incremental improvement by integrating expert feedback into existing methods.
The authors tackled the problem of causal discovery by introducing a Bayesian framework that actively queries expert judgments on edge relations to concentrate the posterior over directed acyclic graphs, resulting in faster posterior concentration and improved recovery of directed effects under tight query budgets.
We propose causal preference elicitation, a Bayesian framework for expert-in-the-loop causal discovery that actively queries local edge relations to concentrate a posterior over directed acyclic graphs (DAGs). From any black-box observational posterior, we model noisy expert judgments with a three-way likelihood over edge existence and direction. Posterior inference uses a flexible particle approximation, and queries are selected by an efficient expected information gain criterion on the expert's categorical response. Experiments on synthetic graphs, protein signaling data, and a human gene perturbation benchmark show faster posterior concentration and improved recovery of directed effects under tight query budgets.