Goal-Oriented Sequential Bayesian Experimental Design for Causal Learning
This addresses the problem of inefficient experimentation in causal learning for researchers, offering a more targeted approach that is particularly beneficial in resource-constrained settings.
The paper tackles the problem of inefficient causal experimental design by introducing GO-CBED, a goal-oriented Bayesian framework that directly maximizes information gain on user-specified causal queries rather than inferring full causal models, resulting in consistent outperformance of existing baselines across various tasks, especially with limited experimental budgets.
We present GO-CBED, a goal-oriented Bayesian framework for sequential causal experimental design. Unlike conventional approaches that select interventions aimed at inferring the full causal model, GO-CBED directly maximizes the expected information gain (EIG) on user-specified causal quantities of interest, enabling more targeted and efficient experimentation. The framework is both non-myopic, optimizing over entire intervention sequences, and goal-oriented, targeting only model aspects relevant to the causal query. To address the intractability of exact EIG computation, we introduce a variational lower bound estimator, optimized jointly through a transformer-based policy network and normalizing flow-based variational posteriors. The resulting policy enables real-time decision-making via an amortized network. We demonstrate that GO-CBED consistently outperforms existing baselines across various causal reasoning and discovery tasks-including synthetic structural causal models and semi-synthetic gene regulatory networks-particularly in settings with limited experimental budgets and complex causal mechanisms. Our results highlight the benefits of aligning experimental design objectives with specific research goals and of forward-looking sequential planning.