Fair Attribute Classification through Latent Space De-biasing
This work is significant for the computer vision community, as it tackles the critical problem of fairness in visual recognition systems by mitigating biases stemming from correlations between target labels and protected attributes.
This paper addresses bias in visual recognition systems where target labels correlate with protected attributes. The authors propose a method using GANs to generate balanced training data by perturbing images in the latent space, resulting in improved fairness and accuracy of target classifiers.
Fairness in visual recognition is becoming a prominent and critical topic of discussion as recognition systems are deployed at scale in the real world. Models trained from data in which target labels are correlated with protected attributes (e.g., gender, race) are known to learn and exploit those correlations. In this work, we introduce a method for training accurate target classifiers while mitigating biases that stem from these correlations. We use GANs to generate realistic-looking images, and perturb these images in the underlying latent space to generate training data that is balanced for each protected attribute. We augment the original dataset with this perturbed generated data, and empirically demonstrate that target classifiers trained on the augmented dataset exhibit a number of both quantitative and qualitative benefits. We conduct a thorough evaluation across multiple target labels and protected attributes in the CelebA dataset, and provide an in-depth analysis and comparison to existing literature in the space.