Preprint: How We Are Advancing Uncertainty Estimation with Deep Learning

By Laura Read on January 20, 2021

Illustration from the preprint illustrating a Mixture Density Network
Illustration from the preprint illustrating a Mixture Density Network

Our Technical Director, Alden Keefe Sampson is a co-author in the academic preprint, Uncertainty Estimation with Deep Learning for Rainfall–Runoff Modelling. The full author list is: Daniel Klotz, Frederik Kratzert, Martin Gauch, Alden Keefe Sampson, Günter Klambauer, Sepp Hochreiter, and Grey Nearing.

Abstract. Deep Learning is becoming an increasingly important way to produce accurate hydrological predictions across a wide range of spatial and temporal scales. Uncertainty estimations are critical for actionable hydrological forecasting, and while standardized community benchmarks are becoming an increasingly important part of hydrological model development and research, similar tools for benchmarking uncertainty estimation are lacking. We establish an uncertainty estimation bench-5marking procedure and present four Deep Learning baselines, out of which three are based on Mixture Density Networks and one is based on Monte Carlo dropout. Additionally, we provide a post-hoc model analysis to put forward some qualitative understanding of the resulting models. Most importantly however, we show that accurate, precise, and reliable uncertainty estimation can be achieved with Deep Learning.