S05

Exploring the frontiers of astrophysics with machine learning: New methods and current challenges

Communauté

Contact : nina.kessler@u-bordeaux.fr

Machine learning (ML) has become an integral part of astrophysical research, transforming how data is processed, analyzed and interpreted. The techniques developed by the community are constantly evolving, offering new opportunities and challenges for many topics including cosmological simulations, the study of high-energy phenomena, star characterization, asteroid detection, automatic recognition of land types on planets, or the prediction of critical parameters for Space Weather.

This session aims to bring together members of the astrophysical ML community, as well as those who are curious about the topic, to exchange on the developed methods and their applications for scientific research. We propose a workshop divided into two parts, the first one will focus on presentations covering ML applications in simulations and data analysis. The second part (~1 hour) will be a round-table to address the challenges that are specific to ML approaches, such as model interpretability, data bias, and the integration of ML with traditional methods. This is an occasion to discuss the strengths, limitations, and potential of current methods in solving astrophysical questions, while also identifying the community’s needs.

SOC : Nina Kessler (LAB), Aurelie Marchaudon (IRAP), Paolo Bianchini (ObAS), Sylvain Breton (INAF), Emeric Bron (OBSPM), David Cornu (OBSPM), Pierre Gratier (LAB), Hugo Vivien (LAM)