# Considering Race a Problem of Transfer Learning

## TL;DR

This paper investigates how race affects transfer learning in facial classification and image synthesis, proposing techniques to improve performance and analyzing domain differences across racial groups.

## Contribution

It introduces a novel approach of considering race as a boundary for transfer learning and demonstrates methods to enhance facial classification and analyze domain shifts in synthesis.

## Key findings

- Transfer learning techniques outperform models trained within the target domain.
- GANs trained on one race show performance drops when tested on another.
- Race-based domain differences are significant in facial image synthesis.

## Abstract

As biometric applications are fielded to serve large population groups, issues of performance differences between individual sub-groups are becoming increasingly important. In this paper we examine cases where we believe race is one such factor. We look in particular at two forms of problem; facial classification and image synthesis. We take the novel approach of considering race as a boundary for transfer learning in both the task (facial classification) and the domain (synthesis over distinct datasets). We demonstrate a series of techniques to improve transfer learning of facial classification; outperforming similar models trained in the target's own domain. We conduct a study to evaluate the performance drop of Generative Adversarial Networks trained to conduct image synthesis, in this process, we produce a new annotation for the Celeb-A dataset by race. These networks are trained solely on one race and tested on another - demonstrating the subsets of the CelebA to be distinct domains for this task.