On Causal and Anticausal Learning

TL;DR

This paper explores how understanding causal relationships can improve function estimation in machine learning, especially under challenges like covariate shift and semi-supervised learning, supported by empirical evidence.

Contribution

It introduces a hypothesis on when semi-supervised learning benefits from causal knowledge and provides empirical validation for this hypothesis.

Findings

Causal knowledge can enhance semi-supervised learning effectiveness.

The paper formulates conditions under which causal understanding aids learning.

Empirical results support the proposed hypothesis.

Abstract

We consider the problem of function estimation in the case where an underlying causal model can be inferred. This has implications for popular scenarios such as covariate shift, concept drift, transfer learning and semi-supervised learning. We argue that causal knowledge may facilitate some approaches for a given problem, and rule out others. In particular, we formulate a hypothesis for when semi-supervised learning can help, and corroborate it with empirical results.