Abstract
No assignee for this patent application has been made.Reporters obtained the following quote from the background information supplied by the inventors:“Federated learning is a machine learning technique that train s a federated learning model across multipledecentralized edge devices or serve rs holding local data samples, without sharing the data samples with acentral s erver. Federated learning provides benefits of privacy preserving machine learni ng and continuouslearning on the edge. However, the performance of federated le arning suffers when the data at the devicesis non-independent and identically d istributed (non-IID). Data augmentation is one approach to addressthe non-IID d ata. Another approach is zone-based federated learning.
NETL
NSTL