Federated Machine Learning and Applications. 3 credits, 3 contact hours

Catalog description

Prerequisites: CS 375 . The increasing availability of data has greatly boosted the power of machine learning (ML). However, modern data generation (e.g., from personal devices, within hospitals) fundamentally changes ML pipelines. Unlike traditional pipelines that use centralized datasets collected from the web to train ML models, these new data modes result in heterogeneous siloed data residing in the devices or organizations that generated it. To make use of this decentralized data, we focus on collaborative and federated ML, which enables secure and trustworthy learning across multiple parties and diverse data sources.