A transformer-based generative model for planetary systems

Numerical calculations of planetary system formation  can provide access to correlations between the properties of planets in the same system. Such correlations can, in return, be used in order to guide and prioritize observational campaigns aiming at discovering some types of planets, like Earth twins. Such numerical simulations are, on the other hand, very demanding in term of computing power. We therefore developed a generative model which is capable of capturing correlations and statistical relationships between planets in the same system. Such a model, trained on the Bern model for planetary system formation, offers the possibility to generate large number of synthetic planetary systems with little computational cost, that can be used, for example, to guide observational campaigns.

Our model is based on the transformer architecture which is well-known to efficiently capture correlations in sequences, and is at the basis of all modern Large Language Models. To assess the validity of the generative model, we perform visual and statistics comparison, as well as a machine learning driven tests. We show using different comparison methods that the properties of systems generated by our model are very similar to the ones of the systems computed directly by the Bern model. We also show that different classifiers cannot distinguish between the directly computed and generated populations, adding confidence that the statistical correlations between planets in the same system are similar. 

Our generative model, which we provide to the community on ai4exoplanets.com, can be used to study a variety of problem like understanding correlations between certain properties of planets in systems, or predicting the composition of a planetary system, given some partial information (e.g. presence of some easier-to-observe planets). Yet, the performances of our generative model rely on the ability of the underlying numerical model, here the Bern model, to accurately represent the actual formation process of planetary system. Our generative model could, on the other hand, very easily be re-trained using as input results of other numerical models provided by the community.

This work has been carried out within the framework of the NCCR PlanetS supported by the Swiss National Science Foundation under grants 51NF40_182901 and 51NF40_205606.