A Bridge over Deep Space: Astronomy and AI
Abstract
In this short poster paper, I summarise experiences gained as a coordinator of the Leibniz Science Campus DiTraRe in the context of interdisciplinary studies with a special focus on collaborating with computer scientists upon application of AI methods. I also present a literature review on the current AI applications use cases in the field of astronomy and astrophysics. The goal of this article is manifold: to computer scientists, it presents the current topics in astronomy where AI is being applied, while to astronomers, it gives a review of the state-of-the-art AI techniques. The paper also aims at encouraging closer and more intense interdisciplinary collaborations.
Full text
A Bridge over Deep Space: Astronomy and AI ANNA JACYSZYN FIZ Karlsruhe – Leibniz Institute for Information Infrastructure [email protected] Abstract ⋆Are you using AI at work often? ⋆Do you consider yourself an AI expert? ⋆Do you think that AI experts should collaborate with AI users more often? Our DiTraRe experience (dimension Exploration and Knowledge Organisation) shows that yes: this collaboration is needed. Below I connect my passion: astronomy, with my current work: applied AI. This literature overview: to computer scientists, presents the current topics in astronomy where AI is being applied, while to astronomers, gives a review of the state-of-the-art AI techniques. About Ania ⋆PhD in astrophysics (3d structure of the Magellanic System based on classical pulsators). ⋆LSC DiTraRe coordinator. ⋆Postdoctoral researcher at FIZ KA Knowledge-Driven AI (KDAI) group. Add contact! Connect on LinkedIn! Research Data Management in Astronomy Publication Culture ⋆Astronomy: a shining example in the fields of research data management (RDM) and open science (OS). ⋆Need for unified RDM recognised early and supported by large agencies (i.e. ESA, NASA). ⋆For decades requirements for making the data public and results open access. ⋆Common standards well developed and deeply established: catalogues, file types, archives, software packages, etc.. Graphics: https: // rubinobservatory. org/ explore/ how-rubin-works/ technology/ data . Virtual Observatory ⋆Globally recognised platform for RDM: Virtual Observatory (VO). ⋆Gathers databases originating from different instruments and creates a FAIR system. ⋆Workflow modelled after W3C, the World Wide Web Consortium. ⋆VO Semantics Working Group: vocabularies and ontologies (symbolic AI). Graphics: IVOA members (https: // ivoa. net/ about/ member-organizations. html ). AI in Astronomy Semantics in Astronomy ⋆VO vocabularies and thesauri are limited. ⋆Advancements needed in the semantics environment. ⋆Recently started: OPAL – Ontology Portal for Astronomy Linked-data. ⋆OPAL will combine existing vocabulaires (like UAT) and FAIRify them, it will also expand them into neighbouring fields. Machine Learning in Astronomy ⋆Different ML methods are deeply incorporated into the field. ⋆ML led to multiple milestone discoveries in astronomy. ⋆Example: large volume, cosmological, gravomagnetohydrodynamical simulations of the Universe – Illustris TNG. ⋆A collection: CAMELS. Cosmology and Astrophysics with MachinE Learning Simulations. Graphics: https: // www. tng-project. org/ , https: // sites. google. com/ site/ aicosmo2019/ . GenAI in Astronomy ⋆Astronomy-specific LLMs have been trained, i.e. AstroLLaMa and astroBERT. ⋆AstroMLab – a group dedicated to advancing the applications of LLMs. ⋆Latest models are capable of winning a gold medal at the International Olympiad on Astronomy and Astrophysics. Graphics: https: // huggingface. co/ UniverseTBD/ astrollama . Towards Neurosymbolic AI in Astronomy ⋆GenAI may provide incorrect results (i.e. hallucinations). ⋆It is profitable to combine symbolic AI + genAI → neurosymbolic AI. ⋆Collaborations between astronomers and computer scientists resulting in neurosymbolic methods are soon to be expected to start influencing the field. Graphics: https: // www. nao. ac. jp/ en/ news/ science/ 2020/ 20200811-subaru. html . Read my paper: https://www.ditrare.de/sites/default/files/ 2025-11/9_camera-ready.pdf