Supervisors

Suitable for

Abstract

An emerging paradigm of machine learning in distributed settings is federated learning. The idea is that data of different agents stays with them and learning happens in a relatively "decentralised" manner. One notable aspect of federated learning is the non-iid nature of the data across different clients. In the first part of the project, the student will identify the technical challenges that this poses for fairness properties of the trained machine learning model. The goal of fair machine learning is to ensure that decisions taken by machine learning systems don't discriminate based on sensitive attributes like race and gender. This part will involve getting familiar with the literature and some experiments or simulations. The second part of the project will involve proposing, implementing and evaluating solutions to address the problem. Further, the student will also have the opportunity to work on proving formal guarantees for their algorithm.

Prerequisites: Good understanding and hands-on experience with machine learning and statistics, proficiency in Python, interest in ML fairness. Students are encouraged to reach out to Naman Goel (naman.goel@cs.ox.ac.uk) to discuss more project ideas related to above topics.