Active Learning, Federated Learning, and Explainable AI (XAI) are core techniques in our urban mobility platform, each contributing uniquely to AI-driven insights: i) Active Learning reduces annotation effort by selecting the most informative data points, enhancing model efficiency in complex, data-sparse environments. ii) (vertical) Federated Learning enables decentralized model training across devices without transferring raw data, preserving privacy while leveraging diverse mobility datasets. iii) XAI makes model decisions transparent and interpretable, especially in spatiotemporal contexts like traffic or transit patterns, fostering trust and informed decision-making. We also use the MAIaaS platform for monitoring and versioning data and models, and to visualize Active Learning and XAI outputs, supporting both technical and stakeholder-facing insights.
