Topical focus: New inference for spatio-temporal mappings of the Universe
Machine Learning and Agentic AI are here to stay and are shaping science and astrophysics in particular. Astrophysical inference from spatio-temporal mappings of the Universe puts robust, scalable and uncertainty-aware inference to a test.
This hackathon aims to further strengthen technical expertise, in particular in Agentic AI, specific ML/astro libraries, and select astrophysics problems in need of ML, as well as foster new connections in scientific AI.
Program:
9:15 - 10:30: SSC (Scientific Software Center) course on Agentic AI and Database generation
10:30 - 10:50: Coffee break
10:50 - 11:30: Lightning talks about ML solutions (e.g. density estimation, SBI, clustering, etc) and astrophysics problems in need of ML
11:30 - 12:15: Follow-up tutorial (specific ML/astro libraries), organisation of project work
12:15 - 13:30: Lunch
13:30 - 15:15: Project work (Hackathon)
15:10 - 15:30: Coffee break
15:30 - 17:00: Project work (Hackathon)
until 18:00: Group presentations (5min talks + 2 min questions)
Lunch and coffee breaks are provided.
Organizers:
Armand Rousselot (armand.rousselot@iwr.uni-heidelberg.de)
Divesh Jain (jain@thphys.uni-heidelberg.de)
Ullrich Köthe (Ullrich.Koethe@iwr.uni-heidelberg.de)
Caroline Heneka (heneka@thphys.uni-heidelberg.de)