Probabilistic Modelling is Sufficient for Causal Inference

TL;DR

This paper argues that probabilistic modelling and inference are sufficient for causal inference, demonstrating how causal questions can be addressed without specialized causal tools or notation.

Contribution

It shows that all causal inference questions can be answered within standard probabilistic frameworks, reinterpreting causal tools as derived from general probabilistic methods.

Findings

Causal questions can be answered with probabilistic models.

Causal tools emerge from standard probabilistic inference.

No need for causal-specific frameworks or notation.

Abstract

Causal inference is a key research area in machine learning, yet confusion reigns over the tools needed to tackle it. There are prevalent claims in the machine learning literature that you need a bespoke causal framework or notation to answer causal questions. In this paper, we want to make it clear that you \emph{can} answer any causal inference question within the realm of probabilistic modelling and inference, without causal-specific tools or notation. Through concrete examples, we demonstrate how causal questions can be tackled by writing down the probability of everything. Lastly, we reinterpret causal tools as emerging from standard probabilistic modelling and inference, elucidating their necessity and utility.