I’m super excited for one of our #NeurIPS #probnum poster tomorrow!

You know the textbook plot for Gaussian Processes, where uncertainty drops as the model adapts to each new datum?

In this paper, we do this iterative update for the (linear algebra) *computation* in the GP. A 🧵

As @pnkraemer announced over on the birdsite, the inaugural Probabilistic Numerics Spring School - probnumschool.org - will take place in Tübingen in late March 2023. I am beyond excited!

Whether you are an interested newbie, or a committed #probnum expert, there’s something here for you: 🧵

#introduction

I do research on probabilistic numerical algorithms, focusing quite heavily on differential equations. I've open-sourced a bunch of tools for probabilistic numerics and differential equations, most of it in #JAX, and I'm not planning on stopping anytime soon ✌️

https://pnkraemer.github.io

https://github.com/pnkraemer

Follow me to hear about the new stuff that's happening in the #ProbNum universe and don't be confused by the occasional #matplotlib-post 🎨

Peter Nicholas Krämer