ConceptFederated Optimization
Federated Optimization
A distributed machine learning paradigm that enables multiple clients to collaboratively train models without sharing raw data. It combines privacy preservation with distributed computation and encompasses various problem classes including convex optimization and variational inequalities.
Timeline (last 90 days)
Sep 29
The more local steps there are, the greater the computation amortized per communication round and the higher the communication efficiency, but client drift errors also accumulate due to differences in client data distributions.
Unverified50%