FEDERATED LEARNING IN A DISAGGREGATED RADIO ACCESS NETWORK: abstract simplified (17696712)

The abstract describes systems and techniques for wireless communications. It explains that a network entity can determine the level of heterogeneity in input data for training a machine learning model. Based on this heterogeneity level, the network entity can determine the period for aggregating the data for training the model. The network entity can then obtain updated model parameters from multiple client devices and combine them to create a combined set of updated model parameters.