Edge Learning for Large-Scale Internet of Things: Task-Oriented Efficient Communications

In Internet-of-Things networks, edge learning for a specific task provides intelligent applications and services. As the network size becomes large, different users may generate distinct datasets. Thus, to suit multiple learning tasks for processing datasets, this paper performs an efficient task-oriented edge learning by jointly accounting for wireless resource allocation and model learning errors. In particular, we first exploit multi-user scheduling to alleviate co-channel interference in dense networks. Then, we perform optimal power allocation in parallel for different learning tasks. Due to the high parallelization of the designed algorithm, extensive experimental results corroborate that multi-user scheduling and task-oriented power allocation benefit more efficient communications for distinct learning tasks compared with the state-of-the-art benchmark algorithms.