Shortcut Learning in Binary Classifier Black Boxes: Applications to Voice Anti-Spoofing and Biometrics

TL;DR

This paper investigates shortcut learning in binary classifiers, especially in voice anti-spoofing and biometrics, proposing a new framework to analyze biases and their effects on classifier behavior.

Contribution

It introduces a novel framework combining intervention and observational methods with linear mixed-effects models to analyze biases in black-box classifiers.

Findings

Effective analysis of dataset biases in voice anti-spoofing and biometrics

Insights into how biases influence classifier decisions beyond error rates

Framework applicable to other domains for bias detection

Abstract

The widespread adoption of deep-learning models in data-driven applications has drawn attention to the potential risks associated with biased datasets and models. Neglected or hidden biases within datasets and models can lead to unexpected results. This study addresses the challenges of dataset bias and explores ``shortcut learning'' or ``Clever Hans effect'' in binary classifiers. We propose a novel framework for analyzing the black-box classifiers and for examining the impact of both training and test data on classifier scores. Our framework incorporates intervention and observational perspectives, employing a linear mixed-effects model for post-hoc analysis. By evaluating classifier performance beyond error rates, we aim to provide insights into biased datasets and offer a comprehensive understanding of their influence on classifier behavior. The effectiveness of our approach is demonstrated through experiments on audio anti-spoofing and speaker verification tasks using both statistical models and deep neural networks. The insights gained from this study have broader implications for tackling biases in other domains and advancing the field of explainable artificial intelligence.