AI4Science

An evening on what the research loop looks like once AI touches every step of it — hypothesis, simulation, experiment. The framing from the hosts is that in a decade "AI for science" will just be called science, and the interesting question is what the next version of the loop actually is.

The stated ground: generative and foundation models for molecules and materials, closing the loop between prediction and experiment, and the gap between benchmark performance and results that hold up at the bench. Technical talks from people building these systems, with real failure modes alongside what works, then time to dig in with peers. Speakers and agenda are still to be announced. Free, at the AI Campus in Wedding, which hosts 80+ AI teams.

Why go: one of the few Berlin rooms where ML researchers and wet-lab scientists are both present. Best for applied scientists and ML engineers working on molecules, materials or simulation.