Federated Learning Algorithms

The implementation of federated learning algorithms in Fed2Tier consists of two components: the training part on the client side and the aggregation part on the server and node side. The training functions are coded in the net_lib.py file located in the node/src directory, while the aggregation functions are located in various files within the algorithms folder at the server/src and node/src directory.

The algorithms currently implemented in Fed2Tier are:

Algorithm	Paper	Server	Node
FedAvg	Communication-Efficient Learning of Deep Networks from Decentralized Data	✅	✅
FedDyn	Federated Learning Based on Dynamic Regularization		✅
Scaffold	SCAFFOLD: Stochastic Controlled Averaging for Federated Learning		✅
FedAdagrad	Adaptive Federated Optimization	✅	✅
FedAdam	Adaptive Federated Optimization	✅	✅
FedYogi	Adaptive Federated Optimization	✅	✅
FedProx	FedProx: Federated Learning with Proximity		✅

—