Operationalizing Complex Causes: A Pragmatic View of Mediation

TL;DR

This paper addresses the challenge of estimating causal effects of complex, non-atomic interventions by formalizing the problem and proposing a two-step prediction and testing method, demonstrated on simulated and real data.

Contribution

It introduces a formal framework and novel methods for causal response estimation under crude interventions, extending causal inference to complex objects.

Findings

Effective estimation of crude intervention effects with limited data

Successful identification of mediators of crude interventions

Validated approach on simulated and real-world examples

Abstract

We examine the problem of causal response estimation for complex objects (e.g., text, images, genomics). In this setting, classical \emph{atomic} interventions are often not available (e.g., changes to characters, pixels, DNA base-pairs). Instead, we only have access to indirect or \emph{crude} interventions (e.g., enrolling in a writing program, modifying a scene, applying a gene therapy). In this work, we formalize this problem and provide an initial solution. Given a collection of candidate mediators, we propose (a) a two-step method for predicting the causal responses of crude interventions; and (b) a testing procedure to identify mediators of crude interventions. We demonstrate, on a range of simulated and real-world-inspired examples, that our approach allows us to efficiently estimate the effect of crude interventions with limited data from new treatment regimes.