📢 Happy to share our work-in-progress paper, which will be presented at the AI4Science workshop of SuperComputing 2024! We developed an online active sampling of simulation inputs to create informative training data based on training loss: arxiv.org/abs/2410.05860

My first technical blog post is out. It outlines an online training framework dedicated to #AI4Science applications. By alleviating disk IO it makes training better ✅, faster ⚡, and cheaper 💸. ltmeyer.github.io/2024/10/15/onl… #Ray #rapidsai #AI4Science #hpc

An utterly shameful decision by ETH Zurich. Last time I checked, collective punishment was still prohibited by the Geneva Convention.

On my way to the Supercomputing conference in Atlanta, US; finishing touches on my slides, going to present our last paper 🥳

It was hard but I did it