Gender Artifacts in Visual Datasets
This work highlights the pervasive gender biases in visual datasets, urging researchers and practitioners to develop robust methods rather than attempting removal, which is incremental in addressing bias mitigation.
The paper investigates gender artifacts—visual cues correlated with gender that are learnable by classifiers and have human-interpretable correlates—in COCO and OpenImages datasets, finding them ubiquitous from low-level to high-level features, and claims that removing these artifacts is largely infeasible.
Gender biases are known to exist within large-scale visual datasets and can be reflected or even amplified in downstream models. Many prior works have proposed methods for mitigating gender biases, often by attempting to remove gender expression information from images. To understand the feasibility and practicality of these approaches, we investigate what $\textit{gender artifacts}$ exist within large-scale visual datasets. We define a $\textit{gender artifact}$ as a visual cue that is correlated with gender, focusing specifically on those cues that are learnable by a modern image classifier and have an interpretable human corollary. Through our analyses, we find that gender artifacts are ubiquitous in the COCO and OpenImages datasets, occurring everywhere from low-level information (e.g., the mean value of the color channels) to the higher-level composition of the image (e.g., pose and location of people). Given the prevalence of gender artifacts, we claim that attempts to remove gender artifacts from such datasets are largely infeasible. Instead, the responsibility lies with researchers and practitioners to be aware that the distribution of images within datasets is highly gendered and hence develop methods which are robust to these distributional shifts across groups.