Navigating Shortcuts, Spurious Correlations, and Confounders: From Origins via Detection to Mitigation

TL;DR

This paper unifies the concept of shortcuts in machine learning, providing a formal definition, connecting related fields, and organizing detection and mitigation approaches to improve model robustness.

Contribution

It introduces a comprehensive taxonomy of shortcut learning, links it to related fields, and classifies datasets, advancing understanding and strategies for addressing shortcuts.

Findings

Provides a formal definition of shortcuts

Organizes existing detection and mitigation methods

Classifies datasets for shortcut learning

Abstract

Shortcuts, also described as Clever Hans behavior, spurious correlations, or confounders, present a significant challenge in machine learning and AI, critically affecting model generalization and robustness. Research in this area, however, remains fragmented across various terminologies, hindering the progress of the field as a whole. Consequently, we introduce a unifying taxonomy of shortcut learning by providing a formal definition of shortcuts and bridging the diverse terms used in the literature. In doing so, we further establish important connections between shortcuts and related fields, including bias, causality, and security, where parallels exist but are rarely discussed. Our taxonomy organizes existing approaches for shortcut detection and mitigation, providing a comprehensive overview of the current state of the field and revealing underexplored areas and open challenges. Moreover, we compile and classify datasets tailored to study shortcut learning. Altogether, this work provides a holistic perspective to deepen understanding and drive the development of more effective strategies for addressing shortcuts in machine learning.