Yannis Theodoridis

Federated learning can match centralized ML performance in bike-sharing demand forecasting, opening the door to privacy-preserving urban mobility solutions.

XGBoost crushes neural nets and statistical methods in EV charging demand forecasting, but federated learning provides a privacy-preserving path forward.

Ditch the GenAI hype for shared mobility forecasting: gradient boosting models, fine-tuned for edge deployment, can outperform them.