Learning
Bits

I gave a presentation at the E-AI Summer Workshop on Products and Services on how we use machine learning to develop our QC system for surface data and to communicate its results. The parts about the statistical evaluation of QC tests and the computation of the probabilistic plausibility have been presented before. The middle part is new, and shows the result of Francesco's internship, where we used Neural Processes for learning to combine rain gauge and radar measurements of precipitation.

The first two papers of the GCOS SwissPhenocam project are now available online. The project evaluates the potential of web cameras for phenological observations, and is conducted jointly with the Ecovision lab of Prof. Wegner at the University of Zürich and the Ecosystem Ecology group of Prof. Gessler at the WSL institute.

Nicola will be presenting our recent progress on using deep learning models for camera-based visibility estimation at this year's CIMO TECO conference.

Our work on photographic visualization of weather forecasts has made it to TV, as part of a segment on AI applications for meteorology :-) Thank you to Daniele Nerini for the presentation.

Our paper on photographic visualization of weather forecasts has been published in Volume 2, Issue 1 of Artificial Intelligence for the Earth Systems (AIES), which is a recently established journal of the American Meteorological Society.