By Phat Diep (phdiep) and Sangyoon Lim (sangyoon)
- This project aims to explore the feasibility and accuracy of classifying individuals' nationalities using Machine Learning models (such as convolutional neural networks). The prevailing models for facial recognition are dominated by Westerners, while Asians are underrepresented, resulting in lesser accuracy in recognizing people of Asian heritage than their Western counterparts (Xiong et al., 2019). Moreover, Asian people represented in such datasets are not distinguished by nationality, which leads to a great loss of data for the development and deployment of country related applications for Asians. The purpose of this research is to provide answers to issues about the efficacy of machine learning models in this kind of categorization work, the morality of putting such a system in place, and the possible uses and consequences of this technology.