Representation Learning with Statistical Independence to Mitigate Bias

TL;DR

This paper introduces a novel adversarial training method for learning unbiased representations in machine learning models, effectively reducing bias related to protected variables while maintaining high predictive performance across various datasets.

Contribution

It proposes a new adversarial loss function that minimizes statistical dependence between learned features and bias variables, improving fairness without sacrificing accuracy.

Findings

Effective bias mitigation demonstrated on synthetic and real datasets.

Achieves superior prediction performance with reduced bias.

Applicable to medical images and gender classification tasks.

Abstract

Presence of bias (in datasets or tasks) is inarguably one of the most critical challenges in machine learning applications that has alluded to pivotal debates in recent years. Such challenges range from spurious associations between variables in medical studies to the bias of race in gender or face recognition systems. Controlling for all types of biases in the dataset curation stage is cumbersome and sometimes impossible. The alternative is to use the available data and build models incorporating fair representation learning. In this paper, we propose such a model based on adversarial training with two competing objectives to learn features that have (1) maximum discriminative power with respect to the task and (2) minimal statistical mean dependence with the protected (bias) variable(s). Our approach does so by incorporating a new adversarial loss function that encourages a vanished correlation between the bias and the learned features. We apply our method to synthetic data, medical images (containing task bias), and a dataset for gender classification (containing dataset bias). Our results show that the learned features by our method not only result in superior prediction performance but also are unbiased. The code is available at https://github.com/QingyuZhao/BR-Net/.