The main contributions of this work can be summarized
as follows:
1. We present a novel hybrid model with discriminative
and generative components for spatiotemporal inferences
about weather.
2. We design and implement a data-driven kernel function
that shapes predictions in accordance with physical
laws.
3. We provide an ecient inference procedure that enables
optimization of the predictive model in accordance
with large-scale phenomena.
4. We evaluate the methods with a set of experiments
that highlight the performance and value of the methodology.