Dynamical Tests of a Deep-Learning Weather Prediction Model — mentioned by Majromax

Fri, 02 Aug 2024 12:33:59 GMT

Although it's obviously difficult to crack open an ML model, they do perform enough computation to have potentially learned something like the dynamical equations for the atmosphere. At the same time, some ML models are surprisingly parsimonious. Graphcast has about 37 million trainable parameters, but its output is a forecast (increment) of six variables on a 37-level, quarter-degree lat/lon grid. That's about 235 million outputs for a single forecast date, so it's safe to conclude that Graphcast cannot memorize its training set. Researchers are also explicitly interested in probing the…

Paper and discussion history. A mention is not an endorsement.

Dynamical Tests of a Deep-Learning Weather Prediction Model · HackerPapers discussions

Dynamical Tests of a Deep-Learning Weather Prediction Model — mentioned by Majromax