Causal Entropy Optimization

TL;DR

This paper introduces Causal Entropy Optimization (CEO), a framework that efficiently optimizes causal effects in unknown causal graphs by integrating structure learning and effect optimization under uncertainty.

Contribution

CEO extends Causal Bayesian Optimization to handle causal graph uncertainty, combining structure learning and effect optimization with an information-theoretic approach.

Findings

CEO outperforms CBO in convergence speed on synthetic models.

The method learns causal structures while optimizing effects.

Joint approach improves over sequential structure learning.

Abstract

We study the problem of globally optimizing the causal effect on a target variable of an unknown causal graph in which interventions can be performed. This problem arises in many areas of science including biology, operations research and healthcare. We propose Causal Entropy Optimization (CEO), a framework that generalizes Causal Bayesian Optimization (CBO) to account for all sources of uncertainty, including the one arising from the causal graph structure. CEO incorporates the causal structure uncertainty both in the surrogate models for the causal effects and in the mechanism used to select interventions via an information-theoretic acquisition function. The resulting algorithm automatically trades-off structure learning and causal effect optimization, while naturally accounting for observation noise. For various synthetic and real-world structural causal models, CEO achieves faster convergence to the global optimum compared with CBO while also learning the graph. Furthermore, our joint approach to structure learning and causal optimization improves upon sequential, structure-learning-first approaches.