Classes Are Not Equal: An Empirical Study on Image Recognition Fairness

TL;DR

This empirical study investigates fairness issues in image recognition, revealing that class disparity stems from representation problems and is influenced by model biases, with data augmentation helping improve fairness.

Contribution

The paper introduces the concept of Model Prediction Bias and demonstrates its role in class fairness disparities in image recognition models.

Findings

Fairness issues are due to representation problems, not classifier bias.

Models show greater prediction bias for harder-to-recognize classes.

Data augmentation and representation learning can improve fairness.

Abstract

In this paper, we present an empirical study on image recognition fairness, i.e., extreme class accuracy disparity on balanced data like ImageNet. We experimentally demonstrate that classes are not equal and the fairness issue is prevalent for image classification models across various datasets, network architectures, and model capacities. Moreover, several intriguing properties of fairness are identified. First, the unfairness lies in problematic representation rather than classifier bias. Second, with the proposed concept of Model Prediction Bias, we investigate the origins of problematic representation during optimization. Our findings reveal that models tend to exhibit greater prediction biases for classes that are more challenging to recognize. It means that more other classes will be confused with harder classes. Then the False Positives (FPs) will dominate the learning in optimization, thus leading to their poor accuracy. Further, we conclude that data augmentation and representation learning algorithms improve overall performance by promoting fairness to some degree in image classification. The Code is available at https://github.com/dvlab-research/Parametric-Contrastive-Learning.