Workshop - XAI Astro

Modern astrophysics increasingly relies on deep learning systems to process large, high-dimensional datasets. However, the inherent black-box nature of many AI models presents a significant barrier to scientific trust. Astrophysicists require more than accurate predictions: they need interpretable insights that are consistent with physical laws and scientific reasoning.

This workshop addresses this bottleneck by exploring Explainable AI (XAI) methods tailored to astrophysical research. Its focus is on the design, evaluation, and implementation of AI architectures engineered for rigorous scientific environments and large-scale astronomical data pipelines.

While the main AIMLSystems conference tracks typically address generalized AI infrastructure or broad algorithmic performance, this workshop focuses on the unique constraints of applying AI to the physical sciences. It aims to bring together machine learning researchers, astrophysicists, and data scientists to discuss how transparent and physically meaningful AI systems can support scientific discovery.

Impact: The novelty of the workshop lies in its interdisciplinary approach. Its expected impact is to provide a roadmap for shifting astrophysics research from using AI as a statistical tool to using it as a validated, interpretable, and transparent engine for theoretical discovery.