Federated edge learning is the derived concept of federated learning that exploits edge computing to train the learning model by orchestrating a set of edge devices with an edge server based on locally distributed data. In edge computing applications, Artificial Intelligence (AI) training models find issues with g privacy, network congestion, and latency while exchanging a huge amount of data.
Federated edge learning owns the capability to tackle the difficulties in conventional AI training models. However, the federated edge learning model requires effectiveness in the aspect of energy consumption and latency. Dynamic joint resource allocation is apposite in diminishing the total energy consumption at edge devices and producing low latency.
In dynamic joint resource allocation, the enhanced system efficiency is obtained by utilizing both communication resource allocation for global parameter aggregation and computation resources allocation for local parameter updation explosively. Federated edge computing with dynamic joint resource allocation improves the system energy efficiency in the potential of low-latency and energy consumption by accurately stabilizing the communication-computation energy tradeoff.