Causal Discovery from Changes

TL;DR

This paper introduces a novel causal discovery method that leverages local changes in data distributions to identify causal structures, providing algorithms to determine equivalence classes and demonstrating effectiveness through simulations.

Contribution

The paper presents a new approach to causal discovery based on detecting local changes, along with algorithms for identifying equivalence classes of causal structures from distribution streams.

Findings

Algorithms successfully identify causal structures in simulated data.

Detection of local changes correlates with accurate causal inference.

Error analysis highlights the method's robustness and limitations.

Abstract

We propose a new method of discovering causal structures, based on the detection of local, spontaneous changes in the underlying data-generating model. We analyze the classes of structures that are equivalent relative to a stream of distributions produced by local changes, and devise algorithms that output graphical representations of these equivalence classes. We present experimental results, using simulated data, and examine the errors associated with detection of changes and recovery of structures.