A View From Somewhere: Human-Centric Face Representations
This work addresses the challenge of human-centric face representation for computer vision and fairness applications, offering a novel dataset and framework to reduce biases and improve interpretability.
The authors tackled the problem of representing human faces in a way that aligns with human perception while avoiding biases from categorical labels, by creating a dataset of 638,180 human judgments of face similarity and learning a continuous embedding space. Their method enables accurate prediction of face similarity and provides interpretable dimensions for human decision-making.
Few datasets contain self-identified sensitive attributes, inferring attributes risks introducing additional biases, and collecting attributes can carry legal risks. Besides, categorical labels can fail to reflect the continuous nature of human phenotypic diversity, making it difficult to compare the similarity between same-labeled faces. To address these issues, we present A View From Somewhere (AVFS) -- a dataset of 638,180 human judgments of face similarity. We demonstrate the utility of AVFS for learning a continuous, low-dimensional embedding space aligned with human perception. Our embedding space, induced under a novel conditional framework, not only enables the accurate prediction of face similarity, but also provides a human-interpretable decomposition of the dimensions used in the human-decision making process, and the importance distinct annotators place on each dimension. We additionally show the practicality of the dimensions for collecting continuous attributes, performing classification, and comparing dataset attribute disparities.