Fostering Trustworthiness of Federated Learning Ecosystem through Realistic Scenarios

The present study thoroughly evaluates the most common blocking challenges faced by federated learning (FL) ecosystem and analyzes existing state-of-the-art solutions. A system adaptation pipeline is designed to enable integration of different AI-based tools in FL system, while training conducted under realistic conditions using a distributed hardware infrastructure. suggested system’s robustness are tested against related tool deployment, data heterogeneity, privacy attacks for multiple tasks types. representative set datasets have been selected cover several validation cases each edge device closely reflect real-world scenarios. presents significant outcomes experiments models’ performance conditions, highlighting potential limitations issues that occurred during process.