Representation Learning with Statistical Independence to Mitigate Bias
This addresses bias issues in ML applications like medical studies and face recognition, offering a method to mitigate bias without requiring perfect dataset curation, though it is incremental as it builds on existing adversarial training approaches.
The paper tackles bias in machine learning by proposing a model that learns features with maximum discriminative power and minimal statistical dependence on protected variables, achieving superior prediction performance and unbiased features on synthetic data, medical images, and a gender classification dataset.
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/.