New endowed chair brings astronomy and AI together. ‘AI can help improve our understanding of the Universe’
AI helps astronomers to understand our Universe better, and that role is set to grow. Therefore, ASTRON and Leiden University are establishing a new research field together. Professor by special appointment Joeri van Leeuwen will develop astrophysics-informed AI, and then use it to uncover new astronomy.
The formation, structure, and evolution of our Universe is driven by many different forces and processes. Gravity, light, and explosive outbursts of matter connect the cosmos over vast scales. ‘We now want to find out how AI can help us understand the laws of nature that shape the Universe,’ says the new professor by special appointment, Van Leeuwen.
New telescopes lead to more and more data
‘The Netherlands hosts radio telescopes that are world leading in astronomical research. These generate more data than entire countries,’ says Violette Impellizzeri, Head of Astronomy and Operations at ASTRON, the Netherlands Institute for Radio Astronomy. ‘Analysing all those images already requires supercomputers and AI, and this challenge will only grow with our future telescopes.’ That is why ASTRON and Leiden University are jointly establishing the new endowed chair.
The team will carry out curiosity-driven research in a number of fields were AI bridges computer science and astronomy. Van Leeuwen: ‘Using the latest advances in AI, we will improve telescope processing, optimise efficiency of AI chips for massive data streams, and increase astrophysical fidelity in AI training. That will help us move the boundaries of the unknown.’
Innovation in both astronomy and computer science
‘To gain maximum insight into our Universe from the latest data and simulations, we need AI that is both fast and understandable. Van Leeuwen is a pioneer of that field,’ says Ignas Snellen, scientific director of Leiden Observatory. Marcello Bonsangue, scientific director of the Leiden Institute of Advanced Computer Science, agrees: ‘Plus, astronomical datasets are public and open, offering excellent testbeds for innovations in computer science.’