AI quest to shed light on distant planets

Astronomers are calling on the international data science community to help find new ways of extracting signals from satellite data of exoplanets - planets beyond the solar system.

The Ariel Data Challenge 2024, which takes its name from the Ariel (Atmospheric Remote-sensing Infrared Exoplanet Large-survey) satellite, will task participants with extracting faint planetary signatures from simulated and noisy data.

Set to launch at the end of the decade, Ariel is the first mission dedicated to understanding the atmospheric chemistry and thermal structures of exoplanets.

Drs Lorenzo Mugnai and Andreas Papageorgiou of Cardiff University’s School of Physics and Astronomy generated the simulated data set for the Challenge using advanced computer models of the Ariel spacecraft and sources that it will observe.

Measuring faint signals from exoplanets’ atmospheres is formidably difficult both in making the observations and analysing the data, due to noise sources such as the “jitter noise” caused by minute spacecraft vibrations, and other disturbances.

The innovative methods uncovered through the Challenge could be applied to the data from the Ariel satellite, giving greater insight into the far reaches of space, the organisers say.

Dr Mugnai said: “As astronomers, we've been dealing with these noise reduction issues for decades. We've used that experience to put together this challenge, and now we're excited to see the fresh ideas and solutions the AI community will come up with.

“The results and insights will help us to be as well-prepared as possible to analyse the data from the Ariel satellite after it is launched.”

Dr Papageorgiou added: “For an ultrasensitive mission such as Ariel, it will be vital to understand the data thoroughly and to be able to identify the true exoplanet signatures in the presence of many unwanted effects from the instruments.

“This means paying as much attention to preparing for the data analysis as to building the satellite itself. The Ariel Data Challenge is part of that process, and will bring new ideas from the worldwide AI community to bear on the exciting challenge of learning about alien worlds.”

Led by University College London (UCL) in collaboration with international partners, the Ariel Data Challenge 2024 was launched on July 31 at NeurIPS 2024, a world-renowned machine learning conference.

Dr Kai Hou (Gordon) Yip, of University College London, Ariel Data Challenge Lead, said: “We are excited to see the innovative solutions that the global data science community can bring to this formidable task.”

The competition is sponsored by the French space agency CNES, Spaceflux, and Blue Sky Space Ltd., and supported by the UK Space Agency.

Dr Caroline Harper, Head of Space Science, UK Space Agency said: “Exoplanets are likely to be more numerous in our galaxy than the stars themselves and the techniques developed through this prestigious competition could help open new windows for us to learn about the composition of their atmospheres, and even their weather.”

More details can be found on the Ariel Data Challenge website and @ArielTelescope.