The Sugeno Integral Used for Federated Learning with Uncertainty for Unbalanced Data

Data is crucial in the digital economy. Many businesses collect and use their data to enhance their performance. However, limited data or low data quality can hinder model development, particularly in dynamic environments. To overcome this, companies collecting similar data may opt to exchange knowledge without sharing their data, due to privacy or legal issues. This is where federated learning comes in. In horizontal federated learning, each client (organization) iteratively improves its model, so that it can be regularly aggregated and shared with all clients participating in the federation for further improvements. In federated averaging, the aggregation mechanism is based on the weighted average and the weights depend on the amount of data available to each client. In this paper, we propose to use a more advanced aggregation mechanism, namely the Sugeno integral. The initial results are promising.