Hackathon: Human Meets AI in Scientific Research Replication

As AI agents become central to scientific discovery at the Center for Decoding the Universe, we need systematic evidence on when they aid replication and when they fail to do so. We will attempt to reproduce key results and analyses from research papers using AI agents, as well as directly with domain-expert scientists, documenting what works and what doesn't.

The day will consist of two morning keynotes, followed by hands-on hacking sessions, and will conclude with individual or team presentations.

We encourage you to join together with your research group, as this activity is particularly beneficial for students and postdoctoral researchers. We especially encourage you to bring a paper to replicate (ideally with available data/code).

Through this hackathon, we are also building community across KIPAC, SLAC, HAI/SDS, and CS while advancing our understanding of AI-enabled science.

Email datascience@stanford.edu from your Stanford account to receive the registration password.

The hackathon is led by a dynamic organizing team: Marcelo Alvarez, Ioana Ciucă, Steven Dillmann, Owen Queen, Kate Storey-Fisher, and Christine Ye.

Advisors: Risa Wechsler, Susan Clark, and Ben Nachman.