Federated Learning (FL) allows distributed clients to collaborate in the training of a machine learning model without sharing any data. This SuperUROP aims to take FL techniques closer to the edge of data collection by developing FL techniques along with the necessary privacy considerations specifically tuned for embedded devices. The goal is to extend the FL methods already developed to create protocols functional for such devices. Additionally, the intention is to implement privacy preserving and communication efficient protocols on top of the FL techniques developed for the embedded devices.

Through SuperUROP, I aim to learn more about research and machine learning. MIT is an amazing place because of the opportunity to be part of high-caliber and meaningful research. That is why I am most excited to grow as a researcher by participating in this project. On the other hand, Federated Learning (FL) is a very relevant and challenging field. While I was exposed to ML in 6.036 or 6.S191, FL is a topic I am eager to explore.