Machine learning for exoplanet detection in high-contrast spectroscopy

Revealing exoplanets by leveraging hidden molecular signatures in cross-correlated spectra with convolutional neural networks

¹ Institute for Particle Physics and Astrophysics, ETH Zürich, Wolfang-Pauli-Strasse 27, 8093 Zürich, Switzerland
² Seminar für Statistik, ETH Zürich, Raemistrasse 101, 8092 Zürich, Switzerland
³ Department of Astronomy, University of Michigan, Ann Arbor, MI 48109, USA
⁴ STAR Institute, University of Liège, 19 Allée du Six Août, 4000 Liège, Belgium
⁵ Département d’Astronomie, Université de Genève, 1290 Versoix, Switzerland

Received: 2 January 2024
Accepted: 23 June 2024

Abstract

Context. The new generation of observatories and instruments (VLT/ERIS, JWST, ELT) motivate the development of robust methods to detect and characterise faint and close-in exoplanets. Molecular mapping and cross-correlation for spectroscopy use molecular templates to isolate a planet’s spectrum from its host star. However, reliance on signal-to-noise ratio metrics can lead to missed discoveries, due to strong assumptions of Gaussian-independent and identically distributed noise.

Aims. We introduce machine learning for cross-correlation spectroscopy (MLCCS). The aim of this method is to leverage weak assumptions on exoplanet characterisation, such as the presence of specific molecules in atmospheres, to improve detection sensitivity for exoplanets.

Methods. The MLCCS methods, including a perceptron and unidimensional convolutional neural networks, operate in the cross-correlated spectral dimension, in which patterns from molecules can be identified. The methods flexibly detect a diversity of planets by taking an agnostic approach towards unknown atmospheric characteristics. The MLCCS approach is implemented to be adaptable for a variety of instruments and modes. We tested this approach on mock datasets of synthetic planets inserted into real noise from SINFONI at the K-band.

Results. The results from MLCCS show outstanding improvements. The outcome on a grid of faint synthetic gas giants shows that for a false discovery rate up to 5%, a perceptron can detect about 26 times the amount of planets compared to an S/N metric. This factor increases up to 77 times with convolutional neural networks, with a statistical sensitivity (completeness) shift from 0.7 to 55.5%. In addition, MLCCS methods show a drastic improvement in detection confidence and conspicuity on imaging spectroscopy.

Conclusions. Once trained, MLCCS methods offer sensitive and rapid detection of exoplanets and their molecular species in the spectral dimension. They handle systematic noise and challenging seeing conditions, can adapt to many spectroscopic instruments and modes, and are versatile regarding planet characteristics, enabling the identification of various planets in archival and future data.