Ali Anwar

These researchers are using MSI for two projects:

• Scalable federated learning: This project seeks to develop new distributed machine learning approaches which can be scaled to millions of edge devices such as cellphones, IoT sensors, etc. The researchers are specifically interested in understanding the impact of data and resource heterogeneity in federated learning at scale. MSI is used for simulating thousands of devices to perform federated learning.
• Model storage: This project's goal is to create an efficient model storage system for machine learning. The researchers are analyzing large amounts of machine learning (ML) models and designing a system to store ML data in a cost-efficient and scalable manner.