Ariel

Welcome to the Ariel Machine Learning Data Challenge. The Ariel Space mission is a European Space Agency mission to be launched in 2029. Ariel will observe the atmospheres of 1000 extrasolar planets - planets around other stars - to determine how they are made, how they evolve and how to put our own Solar System in the galactic context.

Updates

  • 26-Jul-2023: Final leaderboard results announced!
  • 26-Jul-2023: Due to popular demand, we have reopened our leaderboard! You can now submit your solutions with the post-competition leaderboard
  • 19-Apr-2023: Calling all participants! Join our slack channel today! It is easier for us to answer your questions and we will be providing regular updates to the competition as well.
  • 14-Apr-2023: Wish you could have more computing power? See our Resources page on free GPU resources provided by DiRAC.

Timeline

  • 14-April - Challenge begins
  • 14-April - Baseline solution and other documentations released
  • 14-April - Challenge is live!
  • 14-21 April - Beta testing period, things may change without prior notification!
  • 18-June - Invitation to Final Evaluation round
  • 30-June - Winners are informed & announced
  • 18 September - Winning solutions presented at ECML-PKDD 2023 Workshop


Understanding worlds in our Milky Way

Today we know of roughly 5000 exoplanets in our Milky Way galaxy. Given that the first planet was only conclusively discovered in the mid-1990's, this is an impressive achievement. Yet, simple number counting does not tell us much about the nature of these worlds. One of the best ways to understand their formation and evolution histories is to understand the composition of their atmospheres. What's the chemistry, temperatures, cloud coverage, etc? Can we see signs of possible bio-markers in the smaller Earth and super-Earth planets? Since we can't get in-situ measurements (even the closest exoplanet is lightyears away), we rely on remote sensing and interpreting the stellar light that shines through the atmosphere of these planets. Model fitting these atmospheric exoplanet spectra is tricky and requires significant computational time. This is where you can help!


Help us to speed up our model fitting!

ADC 2023 task

Today, our atmospheric models are fit to the data using MCMC type approaches. This is sufficient if your atmospheric forward models are fast to run but convergence becomes problematic if this is not the case. This challenge looks at inverse modelling using machine learning. For more information on why we need your help, we provide more background in the about page and the documentation.


Quite a lot harder than last year

Some of you may have realised that in 2022, we ran a similar challenge at NeurIPS. That challenge was a huge success and over 200 teams participated in solving our inverse modelling task. This year, we are taking it up a notch and are making the training set sparser, the chemistry more non-linear and the planet observations harder. Year by year we are getting closer to our goal of having highly realistic and robust ML models for our hardest cases. If you think you've done well last year, give this one a go!


Apply Now for FREE dirac computing resources

In past competitions some participants did not have the computational resources to be competitive. To alleviate this inequality DiRAC are kindly providing GPU resources to a limited number of participants. See our Resources page on how to apply.


There are prizes

The top three participants will get three registrations to ECML-PKDD 2023 or the monetary equivalent. They will also be invited to present their work at the Ariel Space Mission consortium conference hosted at the European Space Agency.


Many thanks to...

ECML-PKDD 2023 for hosting the data challenge and to the UKRI Science and Technology Funding Council (STFC), specifically STFC Scientific Computing, and Distributed Research Utilising Advanced Computing (DiRAC), the UK Space Agency, the Centre National d'Etudes Spatiales (CNES) and the European Research Council for supporting this effort. Also many thanks to the data challenge team and partnering institutes, see here for some info on the team members and partnering institutes, and of course thanks to the Ariel team for technical support and building the space mission in the first place!

Any questions or something gone wrong? Contact us at: exoai.ucl [at] gmail.com


Ariel h2020 UKSA ESA stfc dirac Spaceflux Europlanet Society UCL MLAnalytics ATI Brandeis CNES INAF NAOJ sapienza sciml ucldisi padua wien