Abstract: Supervised machine learning models are trained on historical data to learn a static mapping between their input and output variables. However, when they are deployed on continuously streamed ...