Application Of Machine Learning Methods to detrend Photometric Data from Exoplanet Transit Surveys

Kurzfassung

This thesis will be questioning about the pertinence of a neural network in the fast analysis of the future data. High variabilities in the light curves of young stars can make the search for young exoplanets difficult, it represents therefore a perfect challenge to be tackled with the help of machine learning. We would like to use machine learning methods, particularly neural networks, on the problem of detecting transiting planets around young variable stars. Machine learning methods are nowadays very popular tools to work with time series and images, and have shown promising results and their capabilities in countless domains. Finally, machine learning relies on a heavy quantity of data and the field of exoplanet search has plenty to offer with successful satellite missions like TESS and Kepler, and in the future with PLATO, that alone represents a perfect motivation of the application of neural networks on the search on young planets around young stars.