Google Cloud Research Credits worth 5.400 Euros awarded to Mohak Chadha for his proposal on “High Performance Serverless Federated Learning”


At the Chair of Computer Architecture and Parallel Systems, Mohak Chadha, Anshul Jindal, and Michael Gerndt are working on the development of a framework called FedLess that leverages serverless computing technologies to enable Federated Learning (FL) across a heterogeneous fabric of Function-as-a-Service Platforms.

The credits for Google Cloud will help to (i) improve and extend Fedless to enable high-performance FL using serverless computing technologies and (ii) to demonstrate the viability of using serverless for FL through a comprehensive comparison of FedLess with traditional stateful FL systems hosted on virtual machines.

Link:

arxiv.org/pdf/2111.03396.pdf