Two-step adversarial debiasing with partial learning -- medical image case-studies

TL;DR

This paper proposes a two-step adversarial debiasing method with partial learning to reduce racial bias in medical image AI models, demonstrating effectiveness on chest X-ray and mammogram datasets while maintaining task accuracy.

Contribution

It introduces a novel two-step adversarial debiasing approach with partial learning specifically designed for medical image analysis.

Findings

Significant reduction in racial bias in models

Preservation of classification performance

Effective on chest X-ray and mammogram datasets

Abstract

The use of artificial intelligence (AI) in healthcare has become a very active research area in the last few years. While significant progress has been made in image classification tasks, only a few AI methods are actually being deployed in hospitals. A major hurdle in actively using clinical AI models currently is the trustworthiness of these models. More often than not, these complex models are black boxes in which promising results are generated. However, when scrutinized, these models begin to reveal implicit biases during the decision making, such as detecting race and having bias towards ethnic groups and subpopulations. In our ongoing study, we develop a two-step adversarial debiasing approach with partial learning that can reduce the racial disparity while preserving the performance of the targeted task. The methodology has been evaluated on two independent medical image case-studies - chest X-ray and mammograms, and showed promises in bias reduction while preserving the targeted performance.