Mary Gates Research Scholar, Autumn 2024
Research Project: Fair RL-FL
Project Description: Bias in Machine Learning (ML) can lead to unfair treatment of certain groups, particularly in areas like healthcare and finance, where disparate outcomes can have life-altering consequences. New training techniques aim to improve fairness while preserving privacy. Federated Learning (FL) is one such approach, allowing models to be trained on data from many devices without centralizing it. Instead of sharing raw data, each device trains a local model and sends model updates (adjustments based on its local data)to a central server, which aggregates them into a global model. This protects privacy while enabling large-scale training, but differences in data quality, representation, or access across devices can reinforce bias, leading to models that work well for some groups but poorly for others. This project tests whether a debiasing system can effectively mitigate bias in FL without sacrificing model performance. To tackle this, I’m adapting a Reinforcement Learning (RL) system, where an agent learns by interacting with an environment and receiving rewards for beneficial actions. The agent evaluates fairness using feedback from client devices and adjusts the central model’s weights before redistributing it for further training. Using fairness metrics and accuracy as its reward signal, the agent continuously refines its strategy, learning how to mitigate bias while preserving performance. Early results suggest this method can reduce bias while maintaining strong model accuracy, highlighting its potential for improving fairness in real world FL systems. If successful, this approach could be applied in areas like medical diagnostics, risk assessment in insurance, and hiring algorithms, where biased models can lead to significant real world harm.
What have you learned throughout your research project?
My project pushed me to work on my planning and problem solving skills in real time, particularly when my experiments didn’t run as expected and optimizations were difficult to achieve. Through many build iterations and experiment configurations I learned how to stay flexible under pressure while utilizing my technical skills. That experience of getting to test my skills under real world pressures was an invaluable learning opportunity for me which helped me grow not just as a researcher, but as a more confident and independent thinker.
I became more comfortable with ambiguity and learned how to troubleshoot problems systematically, rather than getting discouraged when things didn’t work the first, or even fifth, time. I also gained a deeper appreciation for the iterative nature of research. Each failed attempt while frustrating, was still important data and still brought me closer to my goal. And beyond just technical problem solving, I learned how to communicate my findings more clearly, both in written form and in discussions with my mentors, a critical component of any work which seeks to make an actual impact.
What piece of advice do you have for future applicants?
My advice to future Mary Gates Research applicants is to be open to failure and iteration, stay persistent, and don’t be afraid to ask questions or seek feedback. And I really mean that, we’re all here to support each other in improving the world. The people around you want you to succeed. Also, remember that communicating your work clearly is just as important as the work itself.