This $100,000 Project Grant awarded by the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program (CFDA 47.070) aims to develop a resilient next-generation (NextG) network design to support federated learning (FL) over mobile devices.
The key products and services to be delivered under this grant include:
Exploiting serverless computing at the edge to resiliently and efficiently provide machine learning (ML) computing as a service for large-scale heterogeneous mobile devices supporting FL.
Proposing an energy-efficient differential privacy design to protect local training data privacy against inference attacks in FL, by jointly considering gradient quantization, sparsification, and wireless channel noise.
Developing a multi-bit over-the-air computation based spectrum accessing design to improve the spectrum accessing resiliency, network scalability, and efficiency for FL model updates under limited spectrum availability.
The funding will support this collaborative research project at Kennesaw State University Research And Service Foundation, Inc. over the period from January 1, 2025 to December 31, 2026.