Matej Kristan presents HIDRA3/D: Recent advances in deep learning for coastal flood prediction

Climate change is exponentially intensifying extreme coastal floods, which
devastate local communities and economies. These events are anticipated to
intensify in future, thus there is a pressing need to develop early warning
systems to mitigate the damages. I will present our current work of the HIDRA
family of deep neural networks designed by an interdisciplinary team of
computer
vision experts and geophysicists. HIDRA is the first machine-learning method
that outperforms the best geophysics sea surface height prediction model, it is
half a million times faster and predicts the sea surface height for 3 days
ahead
with competitive accuracy. In this talk, I will focus on joint multi-point and
dense prediction capabilities. HIDRA is already used in daily prognosis by the
Slovenian Environment Agency and has been integrated into the European Centre
for Medium-Range Weather Forecasts (CMWF) products set.