Optimal experimental design via Bayesian optimization: active causal structure learning for Gaussian process networks

Abstract

We study the problem of causal discovery through targeted interventions. Starting from few observational measurements, we follow a Bayesian active learning approach to perform those experiments which, in expectation with respect to the current model, are maximally informative about the underlying causal structure. Unlike previous work, we consider the setting of continuous random variables with non-linear functional relationships, modelled with Gaussian process priors. To address the arising problem of choosing from an uncountable set of possible interventions, we propose to use Bayesian optimisation to efficiently maximise a Monte Carlo estimate of the expected information gain.

Working paper. Accepted as a poster at the NeurIPS 2019 workshop, "Do the right thing": machine learning and causal inference for improved decision making. (6 pages + references + appendix)

References in corpus (6)

• Causal Discovery from a Mixture of Experimental and Observational Data
• Causal Discovery from Changes
• Gaussian Process Networks
• Almost Optimal Intervention Sets for Causal Discovery
• ABCD-Strategy: Budgeted Experimental Design for Targeted Causal Structure Discovery
• Probabilistic Active Learning of Functions in Structural Causal Models