Fantastic review work on precipitation nowcasting and of where and how ML can help, by Rachel Prudden and team at Informatics Lab Met Office as well as Shakir Mohamed and Suman Ravuri.

Yeah, I need to figure this out.

New blog post! I discuss probabilistic programming with Pyro for event detection. Informatics Lab rachel-prudden.medium.com/probabilistic-…

What's the coolest/clearest application of graph neural networks?

I'm excited about this direction. Lots of potential in both atmospheric science and machine learning, IMO.

Really enjoying the great suggestions people have had on more research taste exercises/strategies. Curating at the end of colah.github.io/notes/taste/ Ideas from Rachel Prudden, Andy Matuschak, Michael Nielsen (via @shreydjain13), Arthur Allshire, Preetum Nakkiran

New preprint on arXiv! "Stochastic downscaling to chaotic weather regimes using spatially conditioned Gaussian random fields with adaptive covariance" Informatics Lab arxiv.org/abs/2105.08188

Causal networks to the rescue! 🤩Want to disentangle and quantify causal pathways of teleconnections in the climate system? Marlene Kretschmer does a terrific job at explaining all you need to know in our new BAMS paper Rachel Prudden Samantha V Adams @niallhrobinson 🎉 American Meteorological Society

Thank you to a truly amazing set of colleagues whose contributions in so many different ways made this possible: Suman Ravuri Karel Lenc Piotr Mirowski @bsky.social Dmitry Kangin Rachel Prudden @niallhrobinson and many other amazing people. ♥️🚀🥳🤩🎉

It's been a long journey! #rightwhatsnext indeed

A short post announcing our recent publication on nowcasting precipitation! A collaboration between @DeepMind Met Office Informatics Lab medium.com/informatics-la…